functional assessment after peripheral nerve injury · extensor postural, latência do reflexo...

2012

Technical University of Lisbon

Faculty of Human Kinetics

Functional Assessment after Peripheral Nerve Injury

- Kinematic Model of the Hindlimb of the Rat

Thesis submitted in the fulfilment of the requirements for the Degree of Doctor in

Motricidade Humana, speciality of Fisioterapia

SANDRA CRISTINA FERNANDES AMADO

Advisor: Doutor António Prieto Veloso

Co-advisor: Doutora Ana Colette Pereira de Castro Osório Maurício

Jury

President: Reitor da Universidade Técnica de Lisboa

Vogals:

Doutor António José de Almeida Ferreira, Professor Catedrático da Faculdade de Veterinária da Universidade Técnica de Lisboa; Doutor António Prieto Veloso, Professor Catedrático da Faculdade de Motricidade Humana da Universidade Técnica de Lisboa; Doutora Ana Colette Pereira de Castro Osório Maurício, Professora Associada com Agregação do Instituto de Ciências Biomédicas Abel Salazar da Universidade do Porto; Doutora Maria Margarida Marques Rebelo Espanha, Professora Associada da Faculdade de Motricidade Humana da Universidade Técnica de Lisboa; Doutor Artur Severo Proença Varejão, Professor Auxiliar com Agregação da Escola de Ciências Agrárias e Veterinárias da Universidade de Trás-os-Montes e Alto Douro; Doutor Paulo Alexandre Silva Armada da Silva, Professor Auxiliar da Faculdade de Motricidade Humana da Universidade Técnica de Lisboa

The following parts of this thesis have been published or submitted for publication:

(1) Amado S, Simões MJ, Armada da Silva PAS, Luís AL, Shirosaki Y, Lopes MA,

et al. Use of hybrid chitosan membranes and N1E-115 cells for promoting

nerve regeneration in an axonotmesis rat model. [Internet]. Biomaterials. 2008

Nov ;29(33):4409-19.

(2) Amado S, Rodrigues JM, Luís AL, Armada-da-Silva P a S, Vieira M, Gartner

A, et al. Effects of collagen membranes enriched with in vitro-differentiated

N1E-115 cells on rat sciatic nerve regeneration after end-to-end repair.

[Internet]. Journal of neuroengineering and rehabilitation. 2010 Jan; 7:7.

(3) Amado S, Armada-da-Silva P, João F, Mauricio AC, Luis AL, Simões MJ, et al.

The sensitivity of two-dimensional hindlimb joints kinematics analysis in

assessing function in rats after sciatic nerve crush. Behavioural brain research.

2011 (submitted);

(4) S Amado, AL Luis, S Raimondo, M Fornaro, MG Giacobini-Robecchi, S

Geuna, F João, AP Veloso, AC Maurício , PAS Armada-da-Silva. 2012

Neuroscience (submitted)

OTHER PUBLICATIONS

(4) A Gärtner, AC Maurício, T Pereira, PAS Armada-da-Silva, S Amado, AP

Veloso, AL Luís, ASP Varejão, S Geuna (2011). Cellular systems and

biomaterials for nerve regeneration in neurotmesis injuries. In Biomaterials /

Book 3. Editor Prof. Rosario Pignatello. InTech (ISBN 979-953-307-295-0).

(5) João, F., Amado, S., Veloso, A., Armada-da-silva, P., & Maurício, Ana C.

(2010). Anatomical reference frame versus planar analysis: implications for

the kinematics of the rat hindlimb during locomotion. Reviews in the

neurosciences, 21(6), 469.

(6) Simões, M. J., Amado, S., Gärtner, A., Armada-Da-Silva, P. a S., Raimondo,

S., Vieira, M., et al. (2010). Use of chitosan scaffolds for repairing rat sciatic

nerve defects. Italian journal of anatomy and embryology = Archivio italiano di

anatomia ed embriologia, 115(3), 190-210.

ORAL PRESENTATIONS

(7) Armada-da-Silva P, Amado S, Borges M, Simões MJ, João F,

Maurício AC, et al. Nerve Decompression Does Not Improve

Sciatic Function in Chronic Constriction Injury in the Rat. In: 2nd

International Fascia Research Congress. Amsterdam: 2009.

Functional assessment after peripheral nerve injury: Kinematic model of the hindlimb of the rat

vi

(8) Amado S, Veloso A, Armada-da-Silva P, João F, Simões MJ,

Vieira M, et al. Biomechanical and Morphological Approach to

Assess Recovery After Peripheral Nerve Injury in Animal Model.

In: 2nd International Fascia Research Congress. Amsterdam:

2009.

(9) Amado S, Veloso AP, Armada-da-silva P, João F, Luís AL,

Geuna S, et al. Análise biomecâanica da recuperação funcional

na reparação de lesõoes do nervo periférico num modelo

animal. In: 3º Congresso Nacional de Biomecânica da

Sociedade Portuguesa de Biomecânica. Bragança: 2009.

(10) João F, Amado S, Veloso A, Armada-da-Silva P, Mauricio A.

Segmental Kinematic Analysis using a tridimensional

reconstruction of rat hindlimb: comparison between 2D and 3D

joint angles. In: ISB 2010.

vii

Contents

CONTENTS VII

ACKNOWLEDGEMENTS IX

SUMMARY XI

RESUMO XII

CHAPTER1–INTRODUCTIONANDMAINOBJECTIVES 1

CHAPTER2–METHODOLOGICALCONSIDERATIONS 19

CHAPTER3‐USEOFHYBRIDCHITOSANMEMBRANESANDN1E‐115CELLSFOR

PROMOTINGNERVEREGENERATIONINANAXONOTMESISRATMODEL 39

CHAPTER4‐EFFECTSOFCOLLAGENMEMBRANESENRICHEDWITHINVITRO‐

DIFFERENTIATEDN1E‐115CELLSONRATSCIATICNERVEREGENERATIONAFTER

END‐TO‐ENDREPAIR 77

CHAPTER5‐THESENSITIVITYOFTWO‐DIMENSIONALHINDLIMBJOINTS

KINEMATICSANALYSISINASSESSINGFUNCTIONINRATSAFTERSCIATICNERVE

CRUSH 105

CHAPTER6‐THEEFFECTOFACTIVEANDPASSIVEEXERCISEINNERVE

REGENERATIONANDFUNCTIONALRECOVERYAFTERSCIATICNERVECRUSHIN

THERAT 135

CHAPTER7‐GENERALDISCUSSIONANDCONCLUSIONS 165


viii

ix

Acknowledgements

Institutional support and partnership was crucial for the successful development of

scientific work. The experimental work was possible to be performed at the Instituto

Nacional Ricardo Jorge and Faculdade de Medicina Veterinaria de Lisboa. I would

like to gratefully acknowledge the valuable support by Doutor José Manuel Correia

Costa, from Laboratório de Parasitologia, Instituto Nacional de Saúde Dr. Ricardo

Jorge (INSRJ), Porto, Portugal and by Doutora Belmira Carrapiço, from Faculdade

de Medicina Veterinária de Lisboa.

I would like to thanks to the Universidade Técnica de Lisboa (UTL) rector, the

president of scientific committee and the president of Faculdade de Motricidade

Humana, also of Instituto Ciências Biomédicas Abel Salazar – Universidade do

Porto, ICETA-CECA, and Escola Superior de Saúde de Leiria - Instituto Politécnico

de Leiria.

I wish to express my sincere appreciation and gratitude to all of those who have

made this thesis possible. In particular I would like to thank:

Professor António Prieto Veloso, my supervisor, and the head of Neuromechanics

Research Group, for your never-ending patience and generous giving of your time

and skills. I am very grateful for you guiding me through this work.

Professor Ana Colette de Castro Osório Maurício, my co-supervisor, for your

enthusiastic support, your door always being open, for you being a great human. I

have learnt a lot from you. Special thanks also to Henrique…

Professor Paulo Armada-da-Silva, my co-author, for your generous giving of

feedback and input during the work with the research projects.

Professor Ana Lúcia Luís, my co-author, for her microsurgery skills and for all

fruitful discussions and advices regarding science and life in general. Thank you also

to her family, Fernando and Rita.

Professor Jorge Rodrigues, my co-author, for his microsurgery skills and for being

a great example of Human Professional.

Professor Stefano Geuna and his team, my co-authors, for his stereological

techniques that stimulated my interest about nerve morphology.

Professor José Domingues and Ascenção Lopes, my co-authors, for the great

opportunity to work at experimental level with biomaterials, regenerative medicine

and tissue engineering.


x

Maria João Simões, Márcia Vieira, my co-authors, for the patience during the hours

spent at the lab performing the videos.

Filipa João, my co-author, for the enthusiastic interest about rats in silica…!

My former research colleagues at the Neuromechanics group: Professors Filomena

Carnide, Filomena Vieira and Carlos Andrade. Maria Machado (Didi), Vera

Moniz-Pereira, Filipa João, Liliana, Silvia Cabral, Ricardo Matias, Wangdoo Kim

thank you for all the great time we have spent together in the lab and of course

during all of our parties.

Professor Carlos Ferreira, thank you for the “Master thesis” format.

Professor Hugo Gamboa and Neuza Nunes for encouraging the study of pattern

recognition.

To all my patients that understood the need felt to initiate a different way to answer

their questions…. Thank you.

For students and my colleagues at Escola Superior de Saúde de Leiria, a special

thanks to: Ana Roque, José Vital, Luís Eva Ferreira, Sílvia Jorge, Dulce Gomes e

Sónia Pós de Mina, Susana Custódio, Sílvia Gonzaga, Filipa Soares.

My friends and family for the time not spent with them… My parents and sister for

always being there and supporting me in every aspects of life. For your never-ending

encouragement to fully explore my curiosity. My gratitude and my admiration see no

limitations.

To Francisco, for the unconditional love, and Claudio that supported the day-by-day

basis.

My research has been supported by Fundação para a Ciência e Tecnologia (FCT)

from Ministério da Ciência e Ensino Superior (MCES), Portugal, through the financed

grant SFRH/BD/30971/2006 and research project PTDC/CTM/64220/2006. Part of

the work was also supported by PRIN and FIRB grants from the Italian MIUR

(Ministero dell’Istruzione, dell’Università e della Ricerca), by the Regione Piemonte

(Programma di Ricerca Sanitaria Finalizzata)

xi

Summary

Gait analysis is increasingly used on research methodology to assess dynamics

aspects of functional recovery after peripheral nerve injury in the rat model, which

ultimately is the goal of treatment and rehabilitation. In this thesis we studied nerve

regeneration using techniques of molecular and cellular biology. Functional recovery

was evaluated using the sciatic functional index (SFI), the static sciatic index (SSI),

the extensor postural thrust (EPT), the withdrawal reflex latency (WRL) and hindlimb

kinematics. Nerve fiber regeneration was assessed by quantitative stereological

analysis and electron microscopy. From our results, hybrid chitosan membranes after

sciatic nerve crush, either alone or enriched with N1E-115 neural cells, may

represent an effective approach for the improvement of the clinical outcome in

patients receiving peripheral nerve surgery. Collagen membrane, with or without

neural cell enrichment, did not lead to any significant improvement in most of

functional and stereological predictors of nerve regeneration that we have assessed,

with the exception of EPT. Extending the kinematic analysis during walking to the hip

joint improved sensitivity of this functional test. For motor rehabilitation, either active

or passive exercises positively affect sciatic nerve regeneration after a crush injury,

possibly mediated by a direct mechanical effect onto the regenerating nerve.

Keywords: functional assessment, rat, joint kinematics, neurorehabilitation,

peripheral nerve, biomaterials.


xii

Resumo

A crescente utilização da análise da marcha após lesão do nervo periférico no

modelo do rato relaciona-se com a necessidade de avaliar aspectos dinâmicos da

recuperação funcional. Na presente tese estudámos a regeneração do nervo com

utilização de técnicas moleculares e celulares. A recuperação funcional foi avaliada

com uso de Índice de funcionalidade do ciático, índice estático do ciático, reflexo

extensor postural, latência do reflexo flexor e cinemática do membro posterior. A

regeneração da fibra nervosa foi avaliada com técnicas estereológicas e microscopia

electrónica. Dos nossos resultados concluímos que as membranas híbridas de

quitosano após lesão de esmagamento do nervo ciático, com ou sem

enriquecimento de células neurais N1E-115, podem representar uma abordagem

efetiva para a melhoria dos resultados clínicos dos pacientes sujeitos a cirurgia do

nervo periférico. As membranas de colageneo, com ou sem enriquecimento de

células neurais, não repercutiram nenhuma melhoria significativa nos parâmetros

preditores funcionais e estereologicos de regeneração do nervo. Verificámos que a

inclusão da articulação da coxo-femoral na analise cinemática de marcha aumentou

a sensibilidade deste teste funcional. Para a reabilitação motora, quer o exercício

ativo quer passivo influenciou a regeneração do nervo após esmagamento,

possivelmente devido a um efeito mecânico na regeneração do nervo periférico.

Palavras-chave: avaliação funcional, rato, cinemática, neuroreabilitação, nervo

periférico, biomateriais

Chapter 1 – Introduction and Main Objectives

2 Functional Assessement After Peripheral Nerve Injury - kinematic model of the hindlimb of the rat

FMH – Technical University of Lisbon

1 Introduction and main objectives

This thesis was focused on the study of functional recovery after biological therapeutic

strategies for repair of peripheral nerve axonotmesis and neurotmesis injuries with

experimental model of the rat. The available background regarding peripheral nerve

repair suggests that: 1) Biomaterials are important for peripheral nerve repair after

injury with promising functional results; 2) Functional recovery is the golden standard to

ascertained efficacy of interventions and results transposition from in vivo experiments;

3) the description of functional movement of the ankle i.e. ankle kinematics during

stance phase is an accurate method for functional assessment. So, the aims of this

thesis were: 1) to evaluate functional recovery after a moderate i.e axonotemesis and

severe i.e neurotemesis sciatic nerve injury and verify if various types of biomaterials,

would affect functional recovery and the morphology of nerve fibers, with emphases on

evaluation of kinematic analysis of the rat ankle; 2) to improve kinematic model for

assessment of functional recovery after sciatic nerve crush injury; 3) to verify if

therapeutic exercise would induce changes on kinematics of gait and nerve

morphology.

After peripheral nerve injury, its inner capability of repair is a remarkable reality.

Previously, our research group reported in vitro results that highlighted the relevance of

Ca2+ (1; 2) and results leads to a significant research to know which biological element

might contribute to synergistically optimize effects of microsurgical techniques and

improve morphological and functional recovery (3-5). Neurotrophic factors have been

the target of intensive research - their role in nerve regeneration and the way they

influence neural development, survival, outgrowth, and branching (6). Among

neurotrophic factors, neurotrophins have been heavily investigated in nerve

regeneration studies. They include the nerve growth factor (NGF), brain-derived

neurotrophic factor (BDNF), neurotrophin-3 (NT-3), and neurotrophin-4/5 (NT-4/5) (7).

Neurotrophic factors promote a variety of neural responses, including survival and

outgrowth of the motor and sensory nerve fibers, and spinal cord regeneration (8).

However, in vivo responses to neurotrophic factors can vary due to the method of their

delivering. Therefore, the development and use of controlled delivery devices are

required for the study of complex systems.

A multidisciplinary team, including Veterinaries, Engineers, Medical doctors like

neurologists and surgeons, through Experimental Surgery has a crucial role in the

development of biomaterials associated to these cellular systems, and in testing the

surgical techniques that involve their application, always considering animal welfare

Introduction and Main Objectives 3

Sandra Cristina Fernandes Amado

and the most appropriate animal model. Rodents, particularly the rat and the mouse,

have become the most frequently utilized animal models for the study of peripheral

nerve regeneration because of the widespread availability of these animals as well as

the distribution of their nerve trunks which is similar to humans (9). Although other

nerve trunks, especially in the rat forelimb, are getting more and more used for

experimental research (10; 11), the rat sciatic nerve is still the far more employed

experimental model as it provides a nerve trunk with adequate length and space at the

mid-thigh for surgical manipulation and/or introduction of grafts or guides (9). Although

sciatic nerve injuries themselves are rare in humans, this experimental model provides

a very realistic testing bench for lesions involving plurifascicular mixed nerves with

axons of different size and type competing to reach and reinnervate distal targets (9).

Common types of experimentally induced injuries include focal crush or freeze injury

that causes axonal interruption but preserves the connective sheaths (axonotmesis),

complete transection disrupting the whole nerve trunk (neurotmesis) and resection of a

nerve segment inducing a gap of certain length. Several biomaterials developed by our

research group (including PLGA with a novel proportion 90:10 of the two polymers,

poly(L-lactide):poly(glycolide), hybrid chitosan and collagen) or available already in the

market like Neurolac® (made of poly-ε-caprolactone) have been tested associated to

cellular systems to promote nerve regeneration after axonotmesis and neurotmesis

injuries in the sciatic nerve experimental model. The cellular systems that have been

studied in this context include an immortalized neural cell line N1E-115.

One primary cause of poor recovery after long-term denervation is a profound

reduction in the number of axons that successfully regenerate through the deteriorating

intramuscular nerve sheaths. Muscle force capacity is further compromised by the

incomplete recovery of muscle fibres from denervation atrophy. Progressive muscle

atrophy and changes in muscle fibres composition are consequences of peripheral

nervous system injury that interrupts communication between skeletal muscle and

neurons. After peripheral nerve injury, alterations of gait pattern are the most relevant

observation (5; 12). Functional recovery stills the main limitation to achieve results

translation, and the understanding of such limitation is greatly dependent also on

research methods and on therapeutic strategies. Choosing the correct functional assay

in the rat model for peripheral nerve injury (13) and spinal cord injury (14) is a question

of intensive research interest as an experimental model. After peripheral nerve injury

there are neural and mechanical disturbances and the amount of sensory endings that

exist is one of the major limitations to study the functional meaning of recovery and

thus, the biomechanical model has been increasingly used. For most recently

published studies, functional tests consider analysis of voluntary movement and reflex



activity. Gait function represents the integration of motor and sensory function and

several gait parameters were considered with significant relationship for functional

evaluation of experimental rat model, only during the last decade it has been studied

with biomechanics research methods. Previously, we have highlighted the importance

of both functional and morphological methods in this model (5) already reported by

Varejão et al. (15), which make possible to have different levels of complexity to study

the peripheral nerve repair process. In experimental laboratory animal model it is a

great advance to measure observable functional gains, until now impossible to obtain

since they were considered subjective behavioral measurements.

Locomotion, from a mechanical point of view, is characterized by a repetitious

sequence of limb motion to move the body forward while maintaining the stance

balance. There are three basic approaches to analyze gait (16): 1) subdividing the

cycle according to the variations in reciprocal floor contact by the two feet; 2) using the

time and distance qualities of the stride; 3) identifying the functional significance of the

events within the gait cycle and designate these intervals as the functional phases of

gait. According to the variations in reciprocal floor contact by the two feet, as the body

moves forward, one limb serves as mobile source of support while the other limb

advances itself to a new support site. Stance phase designates the entire period during

which the foot is on the ground and begins with initial contact (IC). Swing phase is the

phase of the normal gait cycle during which the foot is off the ground. The swing phase

follows the stance phase and is divided into the initial swing, the midswing, and the

terminal swing stages.

Analysis of the free walking pattern of the rat is the method mostly used for

assessment of motor function through motion analysis. Although locomotor activity in

an open field is a stable behavioural test and may be used as an index of the

behavioural-physiological coping style of an individual rat (Basso, Beatie and

Bresnahan (BBB) Locomotor Rating Scale) (17), the information obtained is qualitative,

range from 0 to 21 and distinguish between locomotor features such as flaccid

paralysis, isolated hind-limb joint movements, weight-supported plantar stepping,

coordination, and details of locomotion (eg, toe clearance, paw position) (18).

Specifically for peripheral nerve injury, it was developed a method to measure a

combination of motor and sensory recovery: Sciatic Functional Index (SFI) (19). They

calculated an index of the functional condition of rat sciatic nerve (SFI) that consists on

expressing as a percentage the difference between the measurements of injured hind

paw parameters and the intact contralateral hind paw parameters obtained from

footprints. The main objective of these methods was to extrapolate the nerve function



recovery considering its action on innervated muscles. After sciatic nerve injury,

footprints reveal differences between the normal and experimental print length with the

last one being the largest. Although SFI is a quantitative method, it is dependent on

the pressure exerted by the foot on the floor, and it is restricted to a point in time, which

limits the information obtained. Automutilation, inversion or eversion deformations often

limit the functional assessment with the use of SFI (20), despite this limitation it is a

widely used parameter because of its reliability (21). Bervar (2000) described an

alternative video analysis of static footprint video analysis (SSI) to assess functional

loss following injury to the rat sciatic nerve, during animal standing or periodic rest on a

flat transparent surface used motion analysis with the utilization of video analysis. SSI

static sciatic index was developed based on the premise that the recovery of muscle

tone after nerve injury is a constituent part of integral nerve and muscle functional

recovery and forces acting on the body i.e. body weight and postural muscle tone

during standing influenced footprint parameters (22). The main difference between SFI

and SSI was that the distance between the tip of the third toe and the posterior margin

of the sole discoloured area (PLF), defined as the print length parameter, was not

considered. The authors considered this parameter the most difficult part of the video

analysis and it was subject to observer’s misinterpretation and unwilling measuring

errors with excessive variability and poor correlation between static video method and

dynamic ink track method. They found better reproducibility, high accuracy, more

precise quantification of the degree of functional loss and there are few non-

measurable footprints with SSI.

Walking pattern is the result of several signs that contributes to the performance of

movement (23). Understanding adaptive changes in motor activity associated with

functional recovery following muscle denervation can be achieved with biomechanical

research methods. In biomechanics, movement is studied in order to understand the

underlying mechanisms involved in the movement or in the acquisition and regulation

of skill. Biomechanics involves mechanical measurements used in conjunction with

biological interpretations (24). Mechanicaly, motion production also depends on the

geometry of bones and muscles, which reflects the moment pattern of motion.

The development of scientific methods of monitoring locomotion is based on

computational and mathematical principles. Biomechanics is based on Newton´s

equation of motion. There are some assumptions about Biomechanics to simplify the

complex musculoskeletal system and eliminate the necessity of quantifying the

changes in mass distribution caused by tissue deformation and movement of bodily

fluids: body segments behave as rigid bodies during movements, and in addition,

segmental mass distribution is similar among members of a particular population (23).



Over the last years, the number of parameters used for ankle joint motion analysis has

been increasing, most likely reflecting advances in computer-based video and motion

analysis. In a first use of ankle joint kinematics in peripheral nerve research, Santos et

al. (25) proposed a 2D geometrical model of the joint using two intersecting lines

joining the knee and ankle and the fifth metatarsal head to the ankle. They

demonstrated an increase in joint angle during the swing phase after a crush injury of

the peroneal nerve. After this early study, Lin et al. (26) examined the angular changes

of the ankle joint during the rat walk after sciatic nerve transection and autografting

using a single parameter: the joint angle at mid-stance. Changes in this angle were

reported as being sensitive to assess functional impairment after sciatic nerve injury

after several months of the sciatic transection, when compared to non-injured controls.

More recently, Varejão et al. (27) contributed significantly to improve ankle joint

kinematic analysis in the rat sciatic nerve model by proposing a more thorough

analysis of ankle motion during the stance phase that takes into account well-defined

time events. These authors measured the angle of the ankle joint at initial (toe) contact

(IC), opposite toe-off (around 20% of the duration of stance phase), heel raise (around

40% of the duration of stance phase), and at toe-off (TO) (27; 28). Using these

measurements, the authors could demonstrate the presence of functional deficit

beyond 8 weeks after sciatic nerve crush, in clear contrast to SFI measures. By this

time of recovery after sciatic crush, SFI measures usually show complete recovery (5;

15).

More recently, the use of digital 2D video analysis of ankle motion in rat peripheral

nerve models also includes the swing phase of walking (29). Also using 2D video

analysis and dedicated software for motion analysis, our group reported recently

measures of both angle and angular velocity of the ankle joint during the stance phase

(5). Angular velocity data were calculated in an attempt to raise the precision of ankle

motion analysis and to increase its power in detecting subtle differences in functional

recovery when testing alternative treatments after sciatic crush (27; 28). We reasoned

that functional deficits during walking in rat nerve models may be masked by the high

redundancy and adaptability of the motor apparatus in response to sensorimotor

alterations (5; 15). From a biomechanical perspective, joint rotational velocity has a

more direct relationship with the forces actuating onto the hindlimb segments and

therefore velocity measures may be better indicators of dysfunction caused by nerve

injury. Moreover, ankle position data is sometimes difficult to interpret, for example in

those cases where ankle joint angle remains unaffected in the weeks immediately after



sciatic nerve crush (27; 28). This shows that measures of ankle joint angle taken at

snapshots during walking may lack sensitivity to assess functional impairment.

Increasing the number of measures may not improve the precision of functional

evaluation. In fact, including more kinematic variables in walking analysis may

otherwise raise uncertainty in interpreting data. For example, Varejão et al. (15) report

that ankle angles at TO were similar before and after 4 weeks of sciatic crush, whereas

ankle angles were clearly affected post-injury at other time points during the stance

phase. Similar inconsistency in kinematic measurements of ankle joint motion in rats

after sciatic crush injury has been also reported by us (30). This latter study was

designed in order to prolong the observation up to 12 weeks. A full functional recovery

was predicted by SFI/SSI, Extensor Postural Trust (EPT) and Withdrawal Reflex

Latency (WRL) but not by all ankle kinematics parameters. Moreover, only two

morphological parameters (myelin thickness/axon diameter ratio and fiber/axon

diameter ratio) returned to normal values. Although these results may reflect

phenomena related to nerve regeneration and end-organ reinnervation, such as motor

axons misdirection (34), they also suggest that kinematic parameters display distinct

ability to demonstrate functional alterations after peripheral nerve injury.

Damage to any portion of the reflex arc, including the sciatic nerve can result in loss or

slowing of the reflex response. Reflex activity refers to the neural activity in which a

particular stimulus, by exciting an afferent nerve, produces a stereotyped, immediate

response of muscle. Considering motor control, it is related with sensory feedback

control. Marshal Hall was the first neurologist who has introduced the term reflex into

biology in the 19th century. He thought of the muscle as reflecting a stimulus as a wall

reflects a ball thrown against it. Reflex was defined as the automatic response of a

muscle or several muscles to a stimulus that excites an afferent nerve. The anatomical

pathway used in a reflex is called the reflex arc and it consists of an afferent nerve, one

or more interneurons within the central nervous system and an efferent nerve. Reflexes

are considered unlearned, rapid, predictable motor responses to a stimulus, and occur

over a highly specific neural pathway called reflex arc. Thalhammer and collaborators,

(36), originally proposed the Extensor Postural Trust (EPT) as a part of the evaluation

of motor function in the rat after sciatic nerve injury. It is classified as a postural reflex

reaction. For this test, the entire body of the rat, excepting the hindlimbs, was wrapped

and supporting the animal by the thorax and lowering the affected hindlimb towards the

platform of a digital balance, elicits the EPT. As the animal is lowered to the platform, it

extends the hindlimb, anticipating the contact made by the distal metatarsus and digits.

The force applied to the platform balance was recorded. Sensory function is usually



assessed with the nociceptive withdrawal reflex (WRL), also called the flexion reflex.

The flexor, or withdrawal, reflex is initiated by a painful stimulus and causes automatic

withdrawal of the endangered body part from the stimulus. It was adapted from the

hotplate test as described by Masters and collaborators in 1993 (35) and involves a

protective response to withdraw the site from the stimulus when cutaneous receptors

(Group III and IV afferents) sense a noxious stimulus. It is a polysynaptic reflex that is

induced by noxious stimulation of the limb and its latency, amplitude, and duration

depends on stimulus intensity (28). Although WRL reflex was originally termed flexion-

withdrawal reflex (37), it has since been shown to involve other muscles besides

flexors (38). The WRL can be variable because it depends on which afferents are

activated by the stimulus and is transmitted over polysynaptic pathways, which means

that the input signal can be modified along its path.

Quantification of the number of myelinated fibers in peripheral nerve is one of the main

indicators of success of peripheral nerve repair. Number, density and diameter of the

nervous fibers are the main variables considered. However, sampling scheme is crucial

to avoid systematic errors of estimation (39) due to, for instance, anisotropy of the

nervous fibres along the nerve and consequent tendency of grouping fibers related with

their diameter (40). The golden standard is the designed-based sampling also

recognized as systematic random sampling scheme. Previously, we have reported that

12 weeks after injury, regenerated nerves have higher mean density and total number

of myelinated axons as lower mean fiber diameter and myelin thickness. Fiber density

and number in crushed nerves is still significantly higher than normal nerves while size

is still significantly lower (5). It has been hypothesized that one explanation could be

the occurrence of a sprouting of more than one growth cone from each severed axon

leading to the presence of an abundance of small regenerating axons that cross the

lesion site and grow to innervate the end organs (46). The primary function of

peripheral nerves is communication. Thus, electrical and/or chemical messages are

passed from neuron to neuron or from neuron to target organ. Curiously, early work on

impulse conduction along peripheral fibers by Erlanger and Gasser (for which they

shared the Nobel Prize in 1942) demonstrated remarkable relationships between the

conduction velocity of the axons and the type of information that was conveyed. In

what concerns nerve morphology, peripheral nerves demonstrate a wide variety of

axonal types, from myelinated axons of 20 microns in diameter, to very fine

nonmyelinated axons as small as 0.2 microns in diameter. Significant correlation and

internal consistency between electrophysiological and morphological parameters (i.e.

conduction velocity and fiber diameter) make available information and guidelines for



parameter selection based upon the specific question being investigated (36). The

largest motor fibers (13-20 um, conducting at velocities of 80 -120 m/s) innervate the

extrafusal fibers of the skeletal muscles, and smaller motor fibers (5-8 um, conducting

at 4-24 m/s) innervate intrafusal muscle fibers. The largest sensory fibers (13-20 µm)

innervate muscle spindles and Golgi tendon organs (both conveying unconscious

proprioceptive information), the next largest (6-12 µm) convey information from

mechanoreceptors in the skin, and the smallest myelinated fibers 1 – 5 µm) convey

information from free nerve endings in the skin, as well as pain, and cold receptors.

Non myelinated peripheral C fibers (0.2 – 1.5 µm) carry information about pain and

warmth.

In our fist study (chapter 3) functional recovery was assessed after different therapeutic

strategies to improve sciatic nerve repair after axonotemesis injury. The experimental

model based on the induction of a crush injury (axonotmesis) in the rat sciatic nerve

provides a very realistic testing bench for lesions involving plurifascicular mixed nerves

with axons of different size and type competing to reach and re-innervate distal targets

(35). This type of injury is thus appropriate to investigate the cellular and molecular

mechanisms of peripheral nerve regeneration, to assess the role of different factors in

the regeneration process (36) and to perform preliminary in vivo testing of biomaterials

that will be useful in tube-guide fabrication for more serious injuries of the peripheral

nerve, such as neurotmesis. Reflex activity, gait function, motion analysis during gait

and nerve morphology were assessed with EPT and WRL, SFI, ankle kinematics and

nerve histomorphometry respectively. Motion pattern of ankle joint was studied

performing 2-D biomechanical analysis (sagital plan) during stance phase of gait. It

was carried out applying a two-segment model of the ankle joint. It was characterized

with four time events: Initial Contact, Opposite Toe-Off, Hell Rise, and Toe-Off as

described above. We brought together two of the more promising recent trends in

nerve regeneration research: 1) local enwrapping of the lesion site of axonotmesis by

means of hybrid chitosan membranes; 2) application of a cell delivery system to

improve local secretion of neurotrophic factors. First, type I, II and III chitosan

membranes were screened by an in vitro assay, in terms of physical, mechanical and

cytocompatibility properties. Then, membranes were evaluated in vivo to assess their

biocompatibility and their effects on nerve fiber regeneration and nerve recovery in a

standardized rat sciatic nerve crush injury model (37). Among the various substances

proposed for the fashioning of nerve conduits, chitosan attracted particular attention

because of its biocompatibility, biodegradability, low toxicity, low cost, enhancement of

wound-healing and antibacterial effects (38). In addition, the potential application of



chitosan in nerve regeneration has been demonstrated both in vitro and in vivo (39).

Chitosan is a partially deacetylated polymer of acetyl glucosamine obtained after the

alkaline deacetylation of chitin (40).

In our second study (chapter 4) different therapeutic strategies were developed to

improve sciatic nerve repair after a different, more severe lesion i.e. neurotmesis with

and without gap. The aim of this study was to verify if rat sciatic nerve regeneration

after end-to-end reconstruction can be improved by seeding in vitro differentiated N1E-

115 neural cells on a type III equine collagen membrane and enwrap the membrane

around the lesion site. Reflex activity, gait function, motion analysis during gait and

nerve morphology were assessed with EPT and WRL, SSI, ankle kinematics and nerve

morphometry respectively. 2-D biomechanical analysis (sagital plan) was carried out

applying a two-segment model of the ankle joint. As in the previous study, we only

considered the stance phase.

In chapter 5, it is described a third group of experiments, that established different

levels of functional dysfunction by evaluating sciatic-crushed rats: 1) in the denervation

phase (one week after injury), 2) in the reinnervated phase: 12 weeks after injury, and

3) sham-operated controls. We considered the entire gait cycle i.e. stance and swing

phases, which were characterized with four time events: Initial Contact, Midstance,

Toe-Off and Midswing. Peak values of joint angle and angular velocity were studied in

both phases: stance and swing. Additionally, all hindlimb joints were studied: ankle,

knee and hip, allowing the study of 2-D motion analysis (sagital plan) and interjoint

coordination. Gait was also characterized in terms of spatiotemporal parameters (gait

velocity, step length, swing and stance phase duration), which allowed having insights

about its mechanical characteristics. An important statistical question was raised with

this study: Increasing the number of ankle motion measures is also not statistically

desirable, particularly if these carries redundancy and lowers sensitivity. Studies in

animal models of peripheral nerve injury are usually constrained by low number of

subjects according to international welfare laws. When kinematic measures are applied

to determine changes at behavioural level, less than 10 animals per group are

commonly used (41-45). Therefore the decision of which kinematic variables should be

used to assess functional recovery in peripheral nerve research should be based on an

evaluation of their sensitivity to detect functional changes of different levels of severity.

A proper selection of kinematic variables would give researchers a tool to monitor

functional recovery after nerve injury and to detect long term functional changes

caused by incomplete recovery or adaptive changes in motor patterns. We performed a



discriminant analysis of the kinematic parameters to verify if 2D joint motion analysis is

highly sensitive and specific to identify functional deficits caused by acute sciatic crush

in the rat. The sensitivity and specificity of a given set or battery of tests may be

evaluated using discriminant analysis. This statistical technique builds a predictive

model to detect membership based on a set of independent variables. This technique

may be applied to evaluate the sensitivity and specificity of a set of testing variables

since it measures the ability of these tests in classifying elements in a specific group.

While assessing sensitivity, discriminant analysis also determines the relative

importance of the independent variables in classifying observations and to discard

variables with little importance in-group segregation. These applications of discriminant

analysis will help to select which of the potentially large number of variables related to

segmental motions during walking will be most important in assessing functional

recovery after a peripheral nerve injury in the rat. Therefore, this study evaluates the

sensitivity and specificity of ankle joint motion kinematic measures in detecting motion

abnormalities as a result of sciatic nerve crush by employing discriminant analysis.

Different levels of functional dysfunction in this study were established by evaluating

sciatic-crushed rats: 1) in the denervation phase (one week after injury), 2) in the

reinnervated phase: 12 weeks after injury, and 3) sham-operated controls.

Our fourth study (chapter 6) aimed to verify if active and passive exercise would elicit

changes in functional recovery after sciatic nerve crush detected by hindlimb

kinematics and nerve morphology. Progressive muscle atrophy and changes in muscle

fibres composition are consequences of peripheral nervous system injury that

interrupts communication between skeletal muscle and neurons. Many strategies have

been used to maintain the muscle activity during the time of reinnervation (46).

Exercise is an important activity in the management of neuromuscular disease. It might

improve return of sensorymotor function. Exercise performed after peripheral nerve

injury acts on muscle properties (histochemical muscle fiber alteration, contractile

properties, enzyme activities, and muscle weight) and nerve properties (axonal

regrowth and myelinization of axons). Most investigations have frequently used the

motorized treadmill in in vivo studies since it offers a controlled and convenient strategy

for testing and training. Twelve weeks after crush injury without exercise protocol,

functional recovery is not full and nerve morphology remains significant different in

control and injured animals as well as ankle joint motion during walk (47). Although

many studies have studied the effects of exercise in functional recovery, biomechanical

evaluation was not considered. Previous studies (48) suggested that functional

reinnervation of hindlimb muscles begins 2 or 3 weeks post sciatic nerve crush in rats



and that overwork of muscle before this period can be harmful (41-45), but the reason

for the negative effect of exercise is still unknown. Therefore, in our study we

performed four groups of animals: two groups of animals started an exercise training

protocol 2 weeks after injury (week 2). Groups (1) sciatic-crushed rats that performed

treadmill walking; (2) sciatic-crushed rats that performed passive muscle stretch (3)

sham-operated controls and (4) sciatic-crushed controls. The exercise groups ended

the program at week-12 post injury. Each rat received 2 weeks of training before

surgery. All rats were subjected to walk/run with no incline at a treadmill speed of 10

m/min continuously, 30 min/day, for 5 days/week during 10 consecutive weeks.

Training was performed on a specially constructed treadmill for rodents developed in

our laboratory with a 10-lane motor-driven conveyer belt with adjustable speed and

inclination. Biomechanical analysis of rat walk was performed every 2 weeks on a

purpose-developed walkway integrating two miniature force plates and a motion

capture system with four high-speed Oqus cameras (Qualysis Systems).



REFERENCES

1. Rodrigues JM, Luís AL, Lobato JV, Pinto MV, Faustino A, Hussain NS, et al.

Intracellular Ca2+ concentration in the N1E-115 neuronal cell line and its use for

peripheric nerve regeneration. [Internet]. Acta médica portuguesa. 2005

;18(5):323-8.

2. Rodrigues JM, Luís AL, Lobato JV, Pinto MV, Lopes MA, Freitas M, et al.

Determination of the intracellular Ca2+ concentration in the N1E-115 neuronal

cell line in perspective of its use for peripheric nerve regeneration. [Internet]. Bio-

medical materials and engineering. 2005 Jan ;15(6):455-65.

3. Luis A, Amado S, Geuna S, Rodrigues J, Simoes M, Santos JD, et al. Long-term

functional and morphological assessment of a standardized rat sciatic nerve

crush injury with a non-serrated clamp [Internet]. Journal of neuroscience

methods. 2007 ;16392–104.

4. Luis AL, Rodrigues JM, Amado S, Veloso AP, Armada-Da-silva PAS, Raimondo

S, et al. PLGA 90/10 and caprolactone biodegradable nerve guides for the

reconstruction of the rat sciatic nerve [Internet]. Microsurgery. 2007 ;27(2):125–

137.

5. Luís AL, Amado S, Geuna S, Rodrigues JM, Simões MJ, Santos JD, et al. Long-

term functional and morphological assessment of a standardized rat sciatic

nerve crush injury with a non-serrated clamp. [Internet]. Journal of neuroscience

methods. 2007 Jun ;163(1):92-104.

6. Schmidt CE, Leach JB. Neural tissue engineering: strategies for repair and

regeneration. [Internet]. Annual review of biomedical engineering. 2003 Jan

;5293-347.

7. Lundborg G, Dahlin L, Danielsen N, Zhao Q. Trophism, tropism, and specificity

in nerve regeneration. [Internet]. Journal of reconstructive microsurgery. 1994

Sep ;10(5):345-54.

8. Frykman GK, McMillan PJ, Yegge S. A review of experimental methods

measuring peripheral nerve regeneration in animals. [Internet]. The Orthopedic

clinics of North America. 1988 Jan ;19(1):209-19.

9. Mackinnon SE, Hudson AR, Hunter DA. Histologic assessment of nerve

regeneration in the rat. [Internet]. Plastic and reconstructive surgery. 1985 Mar

;75(3):384-8.

10. Sinis N, Schaller H-E, Becker ST, Lanaras T, Schulte-Eversum C, Müller H-W,

et al. Cross-chest median nerve transfer: a new model for the evaluation of



nerve regeneration across a 40 mm gap in the rat. [Internet]. Journal of

neuroscience methods. 2006 Sep 30;156(1-2):166-72.

11. Papalia I, Tos P, Scevola A, Raimondo S, Geuna S. The ulnar test: a method for

the quantitative functional assessment of posttraumatic ulnar nerve recovery in

the rat. [Internet]. Journal of neuroscience methods. 2006 Jun 30;154(1-2):198-

203.

12. Varejão ASP, Cabrita AM, Geuna S, Patrício J a, Azevedo HR, Ferreira AJ, et

al. Functional assessment of sciatic nerve recovery: biodegradable poly (DLLA-

epsilon-CL) nerve guide filled with fresh skeletal muscle. [Internet].

Microsurgery. 2003 Jan ;23(4):346-53.

13. Nichols CM, Myckatyn TM, Rickman SR, Fox IK, Hadlock T, Mackinnon SE.

Choosing the correct functional assay: a comprehensive assessment of

functional tests in the rat. [Internet]. Behavioural brain research. 2005 Sep

;163(2):143-58.

14. Sedý J, Urdzíková L, Jendelová P, Syková E. Methods for behavioral testing of

spinal cord injured rats. [Internet]. Neuroscience and biobehavioral reviews.

2008 Jan ;32(3):550-80.

15. Varejão ASP, Cabrita AM, Meek MF, Bulas-Cruz J, Melo-Pinto P, Raimondo S,

et al. Functional and morphological assessment of a standardized rat sciatic

nerve crush injury with a non-serrated clamp. [Internet]. Journal of neurotrauma.

2004 Nov ;21(11):1652-70.

16. Perry J. Gait analysis: Normal and Pathological Function. SLACK Incorporated;

1992.

17. Basso DM, Beattie MS, Bresnahan JC. A sensitive and reliable locomotor rating

scale for open field testing in rats. [Internet]. Journal of neurotrauma. 1995 Feb

;12(1):1-21.

18. Basso DM. Experimental Spinal Cord Injury : Implications of Basic Science

Research for Human Spinal Cord Injury. 2000 ;80(8):808 - 817.

19. Medinaceli L de, Freed WJ, Wyatt RJ. An index of the functional condition of rat

sciatic nerve based on measurements made from walking tracks. [Internet].

Experimental neurology. 1982 Sep ;77(3):634-43.

20. Yu P, Matloub HS, Sanger JR, Narini P. Gait analysis in rats with peripheral

nerve injury. [Internet]. Muscle & nerve. 2001 Feb ;24(2):231-9.

21. Meek MF, Den Dunnen WF, Schakenraad JM, Robinson PH. Long-term

evaluation of functional nerve recovery after reconstruction with a thin-walled

biodegradable poly (DL-lactide-epsilon-caprolactone) nerve guide, using walking



track analysis and electrostimulation tests. [Internet]. Microsurgery. 1999 Jan

;19(5):247-53.

22. Bervar M. Video analysis of standing--an alternative footprint analysis to assess

functional loss following injury to the rat sciatic nerve. [Internet]. Journal of

neuroscience methods. 2000 Oct ;102(2):109-16.

23. Robertson GM, Hamill J, Caldwell G, Kamen G, Whittlesey S. Research

Methods in Biomechanics [Internet]. Human Kinetics Publishers; 2004.

24. Higgins S. Movement as an emergent form: Its structural limits [Internet]. Human

Movement Science. 1985 Jun ;4(2):119-148.

25. Santos PM, Williams SL, Thomas SS. Neuromuscular evaluation using rat gait

analysis. [Internet]. Journal of neuroscience methods. 1995 ;61(1-2):79-84.

26. Lin FM, Pan YC, Hom C, Sabbahi M, Shenaq S. Ankle stance angle: a functional

index for the evaluation of sciatic nerve recovery after complete transection.

[Internet]. Journal of reconstructive microsurgery. 1996 Apr ;12(3):173-7.

27. Varejão ASP, Cabrita AM, Meek MF, Bulas-Cruz J, Gabriel RC, Filipe VM, et al.

Motion of the foot and ankle during the stance phase in rats. [Internet]. Muscle &

nerve. 2002 Nov ;26(5):630-5.

28. Varejão ASP, Cabrita AM, Meek MF, Bulas-Cruz J, Filipe VM, Gabriel RC, et al.

Ankle kinematics to evaluate functional recovery in crushed rat sciatic nerve

[Internet]. Muscle & nerve. 2003 ;27(6):706–714.

29. Ruiter GC de, Spinner RJ, Alaid AO, Koch AJ, Wang H, Malessy MJ a, et al.

Two-dimensional digital video ankle motion analysis for assessment of function

in the rat sciatic nerve model. [Internet]. Journal of the peripheral nervous

system : JPNS. 2007 Sep ;12(3):216-22.

30. Luís AL, Rodrigues JM, Geuna S, Amado S, Simões MJ, Fregnan F, et al.

Neural cell transplantation effects on sciatic nerve regeneration after a

standardized crush injury in the rat. [Internet]. Microsurgery. 2008 Jan

;28(6):458-70.

31. Luís AL, Rodrigues JM, Geuna S, Amado S, Shirosaki Y, Lee JM, et al. Use of

PLGA 90:10 scaffolds enriched with in vitro-differentiated neural cells for

repairing rat sciatic nerve defects. [Internet]. Tissue engineering. Part A. 2008

Jun ;14(6):979-93.

32. Canu M-H, Garnier C. A 3D analysis of fore- and hindlimb motion during

overground and ladder walking: comparison of control and unloaded rats.

[Internet]. Experimental neurology. 2009 Jul ;218(1):98-108.



33. Canu M-H, Garnier C, Lepoutre F-X, Falempin M. A 3D analysis of hindlimb

motion during treadmill locomotion in rats after a 14-day episode of simulated

microgravity. [Internet]. Behavioural brain research. 2005 Feb ;157(2):309-21.

34. Ruiter GCW de, Malessy MJ a, Alaid AO, Spinner RJ, Engelstad JK, Sorenson

EJ, et al. Misdirection of regenerating motor axons after nerve injury and repair

in the rat sciatic nerve model. [Internet]. Experimental neurology. 2008 Jun

;211(2):339-50.

35. Masters DB, Berde CB, Dutta SK, Griggs CT, Hu D, Kupsky W, et al. Prolonged

regional nerve blockade by controlled release of local anesthetic from a

biodegradable polymer matrix. [Internet]. Anesthesiology. 1993 Aug ;79(2):340-

6.

36. Thalhammer JG, Vladimirova M, Bershadsky B, Strichartz GR. Neurologic

Evaluation of the Rat during Sciatic Nerve Block with Lidocaine [Internet].

Anesthesiology. 1995 Apr ;82(4):1013-1025.

37. Sherrington CS. Flexion-reflex of the limb, crossed extension-reflex, and reflex

stepping and standing. [Internet]. The Journal of physiology. 1910 Apr ;40(1-

2):28-121.

38. Schouenborg J, Holmberg H, Weng HR. Functional organization of the

nociceptive withdrawal reflexes. II. Changes of excitability and receptive fields

after spinalization in the rat. [Internet]. Experimental brain research.

Experimentelle Hirnforschung. Expérimentation cérébrale. 1992 Jan ;90(3):469-

78.

39. Geuna S, Gigo-Benato D, Rodrigues A de C. On sampling and sampling errors

in histomorphometry of peripheral nerve fibers. [Internet]. Microsurgery. 2004

Jan ;24(1):72-6.

40. Torch S, Usson Y, Saxod R. Automated morphometric study of human

peripheral nerves by image analysis. [Internet]. Pathology, research and

practice. 1989 Nov ;185(5):567-71.

41. Yuan Y, Zhang P, Yang Y, Wang X, Gu X. The interaction of Schwann cells with

chitosan membranes and fibers in vitro. [Internet]. Biomaterials. 2004 Aug

;25(18):4273-8.

42. Wang W, Itoh S, Matsuda A, Ichinose S, Shinomiya K, Hata Y, et al. Influences

of mechanical properties and permeability on chitosan nano/microfiber mesh

tubes as a scaffold for nerve regeneration. [Internet]. Journal of biomedical

materials research. Part A. 2008 Feb ;84(2):557-66.



43. Chandy T, Sharma CP. Chitosan--as a biomaterial. [Internet]. Biomaterials,

artificial cells, and artificial organs. 1990 Jan ;18(1):1-24.

44. Shirosaki Y, Tsuru K, Hayakawa S, Osaka A, Lopes MA, Santos JD, et al. In

vitro cytocompatibility of MG63 cells on chitosan-organosiloxane hybrid

membranes. [Internet]. Biomaterials. 2005 Feb ;26(5):485-93.

45. Yamaguchi I, Itoh S, Suzuki M, Osaka A, Tanaka J. The chitosan prepared from

crab tendons: II. The chitosan/apatite composites and their application to nerve

regeneration. [Internet]. Biomaterials. 2003 Aug ;24(19):3285-92.

46. Mackinnon SE, Dellon AL, OʼBrien JP. Changes in nerve fiber numbers distal to

a nerve repair in the rat sciatic nerve model. [Internet]. Muscle & nerve. 1991

Nov ;14(11):1116-22.

47. Dellon AL, Mackinnon SE. Selection of the appropriate parameter to measure

neural regeneration. [Internet]. Annals of plastic surgery. 1989 Sep ;23(3):197-

202.

48. Marqueste T, Alliez J-R, Alluin O, Jammes Y, Decherchi P. Neuromuscular

rehabilitation by treadmill running or electrical stimulation after peripheral nerve

injury and repair. [Internet]. Journal of applied physiology (Bethesda, Md. :

1985). 2004 May ;96(5):1988-95.

49. Senel S, McClure SJ. Potential applications of chitosan in veterinary medicine.

[Internet]. Advanced drug delivery reviews. 2004 Jun ;56(10):1467-80.

Chapter 2 – Methodological considerations

20 Functional Assessment after Peripheral Nerve Injury - kinematic model of the hindlimb of the rat


1 Animals

During this chapter we will describe the methods used for the following chapters.

Experiments were performed in rats (Wistar or Sprague-Dawley) weighing 250-300g

(Charles River Laboratories, Barcelona, Spain). Animals were housed for 2 weeks

before entering the experiment and experimental groups were defined with six or eight

animals each depending on the study. Two animals were housed per cage (Makrolon

type 4, Tecniplast, VA, Italy), in a temperature and humidity controlled room with 12-

12h light / dark cycles, and were allowed normal cage activities under standard

laboratory conditions. The animals were fed with standard chow and water ad libitum.

Adequate measures were taken to minimize pain and discomfort taking into account

human endpoints for animal suffering and distress. Moreover, after surgical intervention

cage environment was enriched for all animals with the goal of minimize stress.

All procedures were performed with the approval of the Veterinary Authorities of

Portugal in accordance with the European Communities Council Directive of November

1986 (86/609/EEC).

Handling

The handling was performed to familiarize animals with the experimenter, with the

environment in which the studies would be performed, and with the manipulations

involved in the neurologic evaluation. This familiarization minimizes the stress-

response during the experimental period. The experimental animals were observed for

exploratory activity, and the latency of grooming was everyday monitored.

Chapter 2 – Methodological considerations 21


2 Microsurgical procedures

Our experimental work, concerning the in vivo testing of neurotmesis and axonotmesis

injury and regeneration, was based on the use of Sasco Sprague-Dawley rat sciatic

nerve. Usually, the surgeries are performed under an M-650 operating microscope

(Leica Microsystems, Wetzlar, Germany). Under deep anaesthesia (ketamine 9 mg/100

g; xylazine 1.25 mg/100 g, atropine 0.025 mg/100 body weight, intramuscular), the right

sciatic nerve is exposed through a skin incision extending from the greater trochanter

to the distal mid-half followed by a muscle splitting incision. After nerve mobilisation, a

transection injury is performed (neurotmesis) using straight microsurgical scissors. The

nerve must be injured at a level as low as possible, in general, immediately above the

terminal nerve ramification. For neurotmesis without gap, the nerve is reconstructed

with an end-to-end suture, with two epineural sutures using de 7/0 monofilament nylon.

For axonotmesis we used a standardized clamping procedure that was described in

details in previous works (1-4). After nerve mobilisation, a non-serrated clamp (Institute

of Industrial Electronic and Material Sciences, University of Technology, Vienna,

Austria) exerting a constant force of 54 N, was used for a period of 30 seconds to

create a 3-mm-long crush injury, 10 mm above the bifurcation into tibial and common

peroneal nerves (4; 5). The starting diameter of the sciatic nerve was about 1 mm,

flattening during the crush to 2 mm, giving a final pressure of p9 MPa. The nerves

were kept moist with 37ºC sterile saline solution throughout the surgical intervention.

Finally the skin and subcutaneous tissues are closed with a simple-interrupted suture

of a non-absorbable filament (Synthofil®, Ethicon). An antibiotic (enrofloxacin, Alsir®

2.5 %, 5 mg / kg b.w., subcutaneously) is always administered to prevent any

infections. To prevent autotomy a deterrent substance must be applied to rats’ right

foot (6; 7). All procedures must be performed with the approval of the Veterinary

Authorities of Portugal in accordance with the European Communities Council Directive

of November 1986 (86/609/EEC).



3 Methods for Functional Assessment of

Reinnervation

Biomechanical model - Kinematic analysis

The first step to perform a biomechanical analysis of body motion is the definition of a

mechanical rigid body model that is an idealized form of simplification the structural

differences of bones and represents a system. The definition of the numbers of body

segments is dependent on the movement that we want to analyze. To study ankle joint

movement during locomotion, we defined two rigid bodies: foot and shank.

Considering the computational setting used, to be possible to detect an object/body

moving, it must recognize the body that moves within a recognized and well-defined

area. Therefore, we have to define the 1) segments of the body we want to evaluate; 2)

the reference system;

Motion analysis software provides time-dependent quantitative data, which can be

obtained from stick figures or from volumetric models representing the animal body.

Digital video images record

Animals walked on a Perspex track with length, width and height of respectively 120,

12, and 15 cm (Figure 1). In order to ensure locomotion in a straight direction, the width

of the apparatus was adjusted to the size of the rats during the experiments, and a

darkened cage was connected at the end of the corridor to attract the animals. The

rats’ gait was video recorded at a rate of 100 Hz images per second (JVC GR-

DVL9800, New Jersey, USA). The camera was positioned at the track half-length

where gait velocity was steady, and 1 m distant from the track obtaining a visualization

field of 14 cm wide. Reference system was defined with four points to perform the area:

0.03m x0.015m.

Only walking trials with stance phases lasting between 150 and 400 ms were

considered for analysis, since this corresponds to the normal walking velocity of the rat

(20–60 cm/s) (8; 9).

Digital video images analysis - Two-dimensional joint kinematic analysis

The video images were stored in a computer hard disk for latter analysis using an

appropriate software APAS® (Ariel Performance Analysis System, Ariel Dynamics, San

Diego, USA). Image data were trimmed using APAS-Trimmer: five frames before Initial

contact of the fingers on the floor and five after Toe Off. Data were digitized manually



(APAS-Digitize and -Transform) to perform image representation and filtered with low

pass digital filter at 6 Hz (APAS filter) to determine coordinates for skin landmarks; and

to obtain kinematic parameters of angular displacement and velocity of joint using DLT

(Direct linear transformation) by Abdel-Aziz and Karara (1971). 2-D biomechanical

analysis (sagital plan) was carried out applying a two-segment model of the ankle joint,

adopted from the model firstly developed by Varejão and co-workers (9). Skin

landmarks were tattooed at 3 points: the proximal edge of the tibia, the lateral

malleolus and, the fifth metatarsal head (Figure 2). The definition of the segments foot

and shank was performed manually with digitalization of these points after selecting the

total frames that fulfilled the stance phase (Figure 3). The rats’ ankle angle was

determined using the scalar product between a vector representing the foot and a

vector representing the lower leg. Four complete walking cycles were analysed per rat.

With this model, positive and negative values of position of the ankle joint indicate

dorsiflexion and plantarflexion, respectively. For each stance phase the following time

points were identified (Figure 4): initial contact (IC), opposite toe-off (OT), heel-rise

(HR) and toe-off (TO) (10; 11), and were time normalized for 100% of the stance

phase. The normalized temporal parameters were averaged over all recorded trials.

Angular ankle’s velocity was also determined (negative values correspond to

dorsiflexion).

Figure 1- Track where animals walked, (Perspex 120cm length, 12cm width, and 15cm height)



Figure 2 - Skin landmarks were tattooed at 3 points: the proximal edge of the tibia, the lateral

malleolus and, the fifth metatarsal head

Figure 3 - Video image of the stance phase of rats locomotion

Motion capture - Optoelectronic system

With technical advances in computer science and the continuous development of

mathematical models, biomechanical modeling has improved. Optoelectronic system of

infrared cameras (Oqus-300, Qualisys, Sweden) operating at a frame rate of 200Hz

tracks the motion of small reflective markers placed on the hindlimb using two infra-red

video cameras and has been used to quantify locomotor motion. To obtain three-

dimensional co-ordinate data for a marker, two cameras must record the marker

IC OT HR TOFigure 4 - Time points during stance phase: initial contact (IC), opposite toe-off (OT), heel-rise (HR)

and toe-off (TO)



position in space. Image-processing software identified the marker locations in each

two-dimensional infra-red camera image to compute its three-dimensional location

relative to a calibration plate that was positioned in the data collection corridor. Two

additional cameras can be used to ensure that data from at least two cameras is

always recorded. Prior to recording movements, the cameras must be calibrated by

way of an object with an array of markers whose positions in space are certified to a

known accuracy. Motion capture (MOCAP) allows the assessment of the instantaneous

positions of markers located on the surface of the skin and, thus, a kinematics analysis

of movement. Passive marker based systems use markers coated with a retroreflective

material to reflect light back that is generated near the cameras lens. The camera’s

threshold can be adjusted so only the bright reflective markers will be sampled. We

used an optoelectronic system of six infrared cameras (Oqus-300, Qualisys, Sweden)

operating at a frame rate of 200Hz was used to record the motion of right hindlimb

during the gait cycle. A new corridor was conceptualized and constructed with force

platform system with four load cells (two for vertical force component and two for

anterior-posterior force component) (Figure 5). Animals walked on a Perspex track with

length, width and height of respectively 120, 12 and 15cm. Two darkened cages were

connected at the extremities of the corridor to facilitate walking.

Figure 5 - Set-up of cameras and corridor for motion capture using optoelectronic system

After shaving, seven reflective markers with 2mm diameter were attached to the right

hindlimb at bony prominences (Figure 6): 1) tip of fourth finger, 2) head of fifth

metatarsal, 3) lateral malleolus, 4) lateral knee joint, 5) trochanter major, 6) anterior

superior iliac spine, and 7) ischial tuberosity. The same operator placed all markers

and the rat was maintained static in a similar position to the walking position with the

aim of minimizing the error introduced by the mobility of skin in relation to the bony

references. All rats previously performed two or three conditioning trials to be

familiarized with the corridor. Initial trials are often rejected because rats stop or rise on



their hindlimbs to explore the track. Sometimes another rat was placed inside the cage

to encourage the trial rat to walk along the track toward it. Cameras were positioned to

minimize light reflection artifacts and to allow recording 4-5 consecutive walking cycles,

defined as the time between two consecutive initial ground contacts (IC) of the right

fourth finger. The motion capture space was calibrated regularly using a fixed set of

markers and a wand of known length (20 cm) moved across the recorded field.

Calibration was accepted when the standard deviation of wand’s length measure was

below 0.4 mm.

Figure 6 - Reflective markers with 2mm diameter were attached to the right hindlimb at bony

prominences: (1) tip of fourth finger, (2) head of fifth metatarsal, (3) lateral malleolus, (4) lateral

knee joint, (5) trochanter major, (6) anterior superior iliac spine, and (7) ischial tuberosity and three

non-colinear markers.

Motion analysis - Two-dimensional joint kinematic analysis

In chapters 5 and 6, the kinematic analysis was performed with Visual3D software.

An absolute reference system (ARS), direction of lab co-ordinate, was defined: a right-

handed orthogonal triad <X, Y, Z> fixed in the track ground. Each of the axes is defined

as: +X axis pointing rightward, +Y axis pointing anteriorly and +Z axis pointing upward.

Additionally, a segmental reference system (SRS) is defined. This system uses

Cartesian coordinates fixed to the rigid body and also has clear anatomical meanings

such as proximal-distal, lateral-medial and anterior-posterior.

A gait cycle was defined as the time interval between two consecutive IC of the right

fifth metatarsal. The definition of a static position as a reference frame to define the

position of the segments was conventionally at TO event (Figure 7). Time events (IC

and TO) were detected manually during a first trial by analysing coordinate data from

head of fifth metatarsal marker. IC was defined when the data values became constant,

and TO when data values increased, with visual inspection of the movement. The axis

considered for analysis was that of the direction of locomotion. After the definition of

the event on the first cycle, it was applied a target pattern recognition for others trials of

the same group.



Six walking cycles were analyzed for each animal. Temporal parameters were

normalized to the total duration of the gait cycle. A spline interpolation (performing a

least-squares fit of a 3rd order polynomial to 10 points) and a 2nd order Butterworth

lowpass filter (cut-off frequency of 6Hz, determined by analysis of the difference

residuals between filtered and non-filtered data (12) were applied to the original marker

coordinates data. Joint angle and joint angular velocity were calculated by dot product

and first derivative of joint angle, respectively, between adjacent segments: shank and

foot for ankle joint; shank and thigh for the knee joint; thigh and pelvis for the hip joint.

Figure 7 - Reference frame to define the position of the segments, conventionally at TO event.

Sciatic Functional Index (SFI) and Static Sciatic Index (SSI)

For SFI, animals were tested in a confined walkway measuring 42-cm-long and 8.2-cm-

wide, with a dark shelter at the end. A white paper was placed on the floor of the rat

walking corridor. The hindpaws of the rats were pressed down onto a finger paint-

soaked sponge, and they were then allowed to walk down the corridor leaving its hind

footprints on the paper (Figure 8). Often, several walks were required to obtain clear

print marks of both feet. Prior to any surgical procedure, all rats were trained to walk in

the corridor, and a baseline walking track was recorded. Subsequently, walking tracks

were recorded every week until the week-8 postoperatively and then on weeks 10 and

12 or on weeks 16 and 20 for axonotmesis and neurotemesis injury, respectively.



Several measurements were taken from the footprints (Figure 9): (I) distance from the

heel to the third toe, the print length (PL); (II) distance from the first to the fifth toe, the

toe spread (TS); and (III) distance from the second to the fourth toe, the intermediary

toe spread (ITS). For both dynamic (SFI) and static assessment (SSI), all

measurements were taken from the experimental (E) and normal (N) sides. Prints for

measurements were chosen at the time of walking based on clarity and completeness

at a point when the rat was walking briskly. The mean distances of three

measurements were used to calculate the following factors (dynamic and static):

Toe spread factor (TSF) = (ETS – NTS)/NTS

Intermediate toe spread factor (ITSF) = (EITS – NITS)/NITS

Print length factor (PLF) = (EPL – NPL)/NPL

Where the capital letters E and N indicate injured (experimental) and non-injured side

(normal), respectively.

SFI was calculated as described by (14) according to the Equation 1:

SFI = -38.3 (EPL – NPL)/NPL + 109.5(ETS-NTS)/NTS + 13.3(EIT-NIT)/NIT – 8.8 = (-

38.3 x PLF) + (109.5 x TSF) + (13.3 x ITSF) – 8.8

Equation 1 - SFI

Static footprints were obtained at least during four occasional rest periods. For the

sciatic static index (SSI) only the parameters TS and ITS, were measured (15)

(Equation 2).

SSI = [(108.44 x TSF) + (31.85 x ITSF)] - 5.49

Equation 2 – SSI

Figure 8 - Paint-soaked sponge and corridor where rats leave its hind footprints on the paper

(13).



Figure 9 - Measurements from the footprints: (PL) distance from the heel to the third toe, the print

length; (TS) distance from the first to the fifth toe, the toe spread; and (ITS) distance from the

second to the fourth toe, the intermediary toe spread (13).

For both SFI and SSI, an index score of 0 is considered normal and an index of -100

indicates total impairment. When no footprints were measurable, the index score of -

100 was given (11). Reproducible walking tracks could be measured from all rats. In

each walking track three footprints were analysed by a single observer, and the

average of the measurements was used in SFI calculations.

Extensor Postural Thrust (EPT) - Motor reflex function

The Extensor Postural Trust was originally proposed by Thalhammer and collaborators,

(17) as a part of the neurological recovery evaluation in the rat after sciatic nerve injury.

For EPT test, the entire body of the rat, except the hindlimbs, was wrapped in a

surgical towel and supported by the thorax (Figure 10). The affected hindlimb was then

lowered towards the platform of a digital balance (model PLS 510-3, Kern & Sohn

GmbH, Kern, Germany) to elicit the EPT. As the animal was lowered over the platform,

it extended the hindlimb, anticipating the contact made by the distal metatarsus and

digits. The force in grams applied to the digital platform balance was recorded (digital

scale range 0-500 g). The reduction in this force, representing reduced extensor

muscle tone, was considered a deficit of motor function. The same procedure was

applied to the contra-lateral, unaffected limb. The affected and normal limbs were

tested 3 times, with an interval of 2 minutes between consecutive tests, and the three

values were averaged to obtain a final result. The normal (unaffected limb) EPT

(NEPT) and experimental EPT (EEPT) values were incorporated into an equation

Equation (11) to derive the percentage of functional deficit, as described in the

literature by (16):

% Motor deficit = [(NEPT – EEPT)/NEPT] x 100

Equation (11) - EPT

Normal(N) Experimental(E)



Figure 10 - Extensor Postural Thrust (13)

Withdrawal Reflex Latency - Nociception

The rat was wrapped in a surgical towel above its waist and then positioned to stand

with the affected hindpaw on a hotplate at 56ºC (Figure 12) (model 35-D; IITC Life

Science Instruments, Woodland Hill, CA). WRL is defined as the time elapsed from the

onset of hotplate contact to withdrawal of the hindpaw (Figure 11) and measured with a

stopwatch. Normal rats withdraw their paws from the hotplate within 4 seconds or less

(18). The affected limbs were tested three times, with an interval of 2 minutes between

consecutive tests to prevent sensitization, and the three latency times were averaged

to obtain a final result (19; 20). The cut off time for heat stimulation was set at 12

seconds to avoid skin damage to the foot (3; 21).

Figure 11 – WRL test (13).



Figure 12 - Hotplate for WRL test (13)



4 Morphological Evaluation

Design-based quantitative morphology and electron microscopy

After the follow-up time, rats were anaesthetised and a 10-mm-long segment of the

sciatic nerve that included the injured portion was collected, fixed, and prepared for

morphological analysis and histomorphometry of myelinated nerve fibers. A 10-mm

segment of uninjured sciatic nerve was also withdrawn from the control animals.

Immediately after collecting the nerve, rats were euthanized through an intracardiac

injection of 5% sodium pentobarbital (Eutasil®). Sciatic nerve samples were immersed

immediately in a fixation solution, containing 2.5% purified glutaraldehyde and 0.5%

saccarose in 0.1M Sorensen phosphate buffer for 6-8 hours. Specimens were then

washed in a solution containing 1.5% saccarose in 0.1M Sorensen phosphate buffer,

post-fixed in 2% osmium tetroxide, dehydrated and embedded in Glauerts' embedding

mixture of resins consisting in equal parts of Araldite M and the Araldite Härter, HY 964

(Merck, Darmstad, Germany), to which was added 1-2% of the accelerator 964, DY

064 (Merck, Darmstad, Germany). The plasticizer dibutyl phthalate was added in a

quantity of 0.5% (19; 20). Series of 2-µm thick semi-thin transverse sections were cut

using a Leica Ultracut UCT ultramicrotome (Leica Microsystems, Wetzlar, Germany)

and stained by Toluidine blue for 2-3 minutes for high resolution light microscopy

examination. In each nerve, histomorphometry was conducted using a DM4000B

microscope equipped with a DFC320 digital camera and an IM50 image manager

system (Leica Microsystems, Wetzlar, Germany). This system reproduced microscopic

images (obtained through a 100x oil-immersion objective) on the computer monitor at a

magnification adjusted by a digital zoom. The final magnification was 6600x enabling

accurate identification and morphometry analysis of myelinated nerve fibers. One semi-

thin section from each nerve was randomly selected and used for the morpho-

quantitative analysis. The total cross-sectional area of the nerve was measured and

sampling fields were then randomly selected using a protocol previously described (3;

21). Briefly, cross-sectional area of the nerve is divided into various equal geometric

fields (usually >15) and then each of these are divided into smaller fields. The first

sampling field is randomly selected and then the selection of the next fields was

defined through a systematic “jump” process. Possible "edge effects" (i.e. counting a

fibre more than once, especially for larger fibres that appear in more than one sampling

field) were compensated by employing a two-dimensional dissector procedure, which is

based on sampling the "tops" of fibers (22). Briefly, considering a two-dimensional

observational field where direction (North/South) is defined and the North top of each



fiber is marked. The top of each fiber represents the point of the shape of the fibre that

first intersects the observational field, and since it appears only once, therefore it will be

counted only once.

Mean fiber density was calculated by dividing the total number of nerve fibers within the

sampling field by its area (N/mm2). Total fibers number (N) was then estimated by

multiplying the mean fiber density by the total cross-sectional area of the whole nerve

cross section assuming a uniform distribution of nerve fibers across the entire section.

Fiber and axon area were measured and the circle-fitting diameter of fiber (D) and axon

(d) were calculated. These data were used to calculate myelin thickness [(D-d)/2],

myelin thickness/axon diameter ratio [(D-d)/2d], and fiber/axon diameter ratio (D/d).

The precision of the histomorphometry methods was evaluated by calculating the

coefficient of error (CE). Regarding quantitative estimates of fiber number, the CE(n)

was obtained as follows (23; 24):

'

1)(

QnCE

Equation 4 - quantitative estimates of fiber number

Where Q' is the number of counted fibers in all dissectors.

For size estimates, the coefficient of error was estimated as follows (22):

Mean

SEMzCE )(

Equation 5 - coefficient of error

Where SEM = standard error of the mean.

The sampling scheme was designed in order to keep the CE below 0.10, which

assures enough accuracy for neuromorphological studies (23; 24).

Transmission electron microscopy

The immunohistochemical technique is based on the use of antibodies that bind

specifically to certain cell antigen and thus became visible by fluorescence microscopy

or confocal laser. For this it is necessary to use certain fluorophores or fluorescent

probes to detect the antigen-antibody complexes. This technique allowed detection of

the axon regeneration and possible migration of Schwann cells within the guide tubes

or in biomaterials during the regeneration of peripheral nerve (26). By

immunhistochemistry, the antibodies used are anti-NF-200kd (antiprotein 200kd

neurofilament) and anti-GFAP (anti-glial protein). The first antibody will allow for tracing

of regenerating axons and the second, the possible migration of Schwann cells within

the guide tubes (27). Nowadays there are a number of antigens available, antigen-

antibody affinity, antibody types and methods of assessment and detection. However, it



is necessary that the method is assessed for each particular situation. For optical

microscopy, is usually used staining hematoxylin-eosin (28).

For transmission electron microscopy, ultra-thin sections were cut by means of the

same ultramicrotome and stained with saturated aqueous solution of uranyl acetate

and lead citrate. Ultra-thin sections were analyzed using a JEM-1010 transmission

electron microscope (JEOL, Tokyo, Japan) equipped with a Mega-View-III digital

camera and a Soft-Imaging-System (SIS, Münster, Germany) for the computerized

acquisition of the images.

Scanning electron microscope analysis

The surface morphology of the chitosan membranes was observed under a scanning

electron microscope (SEM; JEOL JSM 6301F) equipped with x-ray energy dispersive

spectroscopy (EDX) microanalysis capability (Voyager XRMA System, Noran

Instruments).

Cell culture, intracellular calcium concentration ([Ca2+]i) measurements and cell

adherence assays

N1E-115 cell line is a clone of cells derived from mouse neuroblastoma C-1300 [58]

and retains numerous biochemical, physiological, and morphological properties of

differentiated neuronal cells in culture [59]. N1E-115 neuronal cells were cultured in

poly-l-lisine coated Petri dishes (around 2 x 106 cells) on 2 x 2 cm chitosan fragments

(chitosan type I, type II and type III) at 37ºC, 5% CO2 in a humidified atmosphere

(Nuaire). Maintenance medium was 89.8% Dulbecco’s Modified Eagle’s Medium

(DMEM) supplemented with glutamine (GlutaMAX; Gibco), 10% fetal bovine serum

(FBS, Sigma), and 0.1% penicillin/streptomycin (100000 U/ml penicillin, 10 mg/ml

streptomycin; Sigma) and with 0.1% β-amphoterrycin (250 μg/ml, Sigma). The culture

medium was changed every 48 hours and the cells were observed daily in an inverted

microscope. Before surgery, once N1E-115 cell culture reached approximately 80%

confluence, cells were supplied with differentiation medium containing DMSO. The

differentiation medium was composed by 95.8% Dulbecco’s Modified Eagle’s Medium

(DMEM) supplemented with glutamine (GlutaMAX; Gibco), 2.5% FBS, 0.1%

penicillin/streptomycin (100000 U/ml penicillin, 10 mg/ml streptomycin; Sigma), 0.1% β-

amphoterrycin (250 μg/ml, Sigma), and 1.5% DMSO (Sigma). Cell culture viability was

assessed by measuring intracellular free calcium concentration ([Ca2+]i). The [Ca2+]i



was measured in Fura-2-AM-loaded cells, through dual wavelength spectrofluorometry

as previously described (22; 26). [Ca2+]i was determined in N1E-115 cell culture before

differentiation and 24, 48 and 72 hours after the onset of DMSO-induced differentiation,

in order to determine the best period of neural differentiation, before the [Ca2+]I rise and

the initiation of the apoptosis process.



REFERENCES




methods. 2007 Jun ;163(1):92-104.




137.

3. Varejão ASP, Cabrita AM, Meek MF, Fornaro M, Geuna S, Giacobini-Robecchi

MG. Morphology of nerve fiber regeneration along a biodegradable poly (DLLA-

epsilon-CL) nerve guide filled with fresh skeletal muscle. [Internet]. Microsurgery.

2003 Jan ;23(4):338-45.

4. Varejão AS, Melo-Pinto P, Meek MF, Filipe VM, Bulas-Cruz J. Methods for the

experimental functional assessment of rat sciatic nerve regeneration. [Internet].

Neurological research. 2004 Mar ;26(2):186-94.

5. Beer GM, Steurer J, Meyer VE. Standardizing nerve crushes with a non-serrated

clamp. [Internet]. Journal of reconstructive microsurgery. 2001 Oct ;17(7):531-4.

6. Sporel-Ozakat RE, Edwards PM, Hepgul KT, Savas A, Gispen WH. A simple

method for reducing autotomy in rats after peripheral nerve lesions. [Internet].

Journal of neuroscience methods. 1991 Feb ;36(2-3):263-5.

7. Kerns JM, Braverman B, Mathew A, Lucchinetti C, Ivankovich AD. A comparison

of cryoprobe and crush lesions in the rat sciatic nerve. [Internet]. Pain. 1991 Oct

;47(1):31-9.

8. Varejão ASP, Cabrita AM, Patricio JA, Bulas-Cruz J, Gabriel RC, Melo-Pinto P,

et al. Functional assessment of peripheral nerve recovery in the rat: gait

kinematics [Internet]. Microsurgery. 2001 ;21(8):383–388.






nerve. 2002 Nov ;26(5):630-5.

11. Dijkstra JR, Meek MF, Robinson PH, Gramsbergen A. Methods to evaluate

functional nerve recovery in adult rats: walking track analysis, video analysis and



the withdrawal reflex. [Internet]. Journal of neuroscience methods. 2000 Mar

;96(2):89-96.

12. Winter DA. Biomechanics and Motor Control of Human Movement [Internet].

Wiley; 2004.

13. Luis AL. Reparação de lesões do nervo periférico. 2008 Doctoral Thesis;

ICBAS-UP.

14. Bain JR, Mackinnon SE, Hunter DA. Functional evaluation of complete sciatic,

peroneal, and posterior tibial nerve lesions in the rat. [Internet]. Plastic and

reconstructive surgery. 1989 Jan ;83(1):129-38.



neuroscience methods. 2000 Oct ;102(2):109-16.

16. Koka R, Hadlock TA. Quantification of functional recovery following rat sciatic

nerve transection. [Internet]. Experimental neurology. 2001 Mar ;168(1):192-5.




18. Hu D, Hu R, Berde CB. Neurologic evaluation of infant and adult rats before and

after sciatic nerve blockade. [Internet]. Anesthesiology. 1997 Apr ;86(4):957-65.

19. Shir Y, Campbell JN, Raja SN, Seltzer Z. The correlation between dietary soy

phytoestrogens and neuropathic pain behavior in rats after partial denervation.

[Internet]. Anesthesia and analgesia. 2002 Feb ;94(2):421-6.

20. Campbell JN. Nerve lesions and the generation of pain. [Internet]. Muscle &

nerve. 2001 Oct ;24(10):1261-73.

21. Varejão ASP, Cabrita AM, Geuna S, Patrício J a, Azevedo HR, Ferreira AJ, et al.

Functional assessment of sciatic nerve recovery: biodegradable poly (DLLA-


2003 Jan ;23(4):346-53.

22. Geuna S, Tos P, Battiston B, Guglielmone R. Verification of the two-dimensional

disector, a method for the unbiased estimation of density and number of

myelinated nerve fibers in peripheral nerves. [Internet]. Annals of anatomy =

Anatomischer Anzeiger : official organ of the Anatomische Gesellschaft. 2000

Jan ;182(1):23-34.



Jan ;24(1):72-6.



24. Schmitz C. Variation of fractionator estimates and its prediction. [Internet].

Anatomy and embryology. 1998 Nov ;198(5):371-97.

25. Di Scipio F, Raimondo S, Tos P, Geuna S. A simple protocol for paraffin-

embedded myelin sheath staining with osmium tetroxide for light microscope

observation. [Internet]. Microscopy research and technique. 2008 Jul ;71(7):497-

502.

26. Geuna S, Tos P, Guglielmone R, Battiston B, Giacobini-Robecchi MG.

Methodological issues in size estimation of myelinated nerve fibers in peripheral

nerves. [Internet]. Anatomy and embryology. 2001 Jul ;204(1):1-10.

27. Raimondo S, Fornaro M, Di Scipio F, Ronchi G, Giacobini-Robecchi MG, Geuna

S. Chapter 5: Methods and protocols in peripheral nerve regeneration

experimental research: part II-morphological techniques. [Internet]. International

review of neurobiology. 2009 Jan ;8781-103.




;28(6):458-70.

Chapter 3 - Use of hybrid chitosan membranes and

N1E-115 cells for promoting nerve regeneration in an

axonotmesis rat model

Amado S, Simões MJ, Armada da Silva PAS, Luís AL, Shirosaki Y, Lopes MA, et al.

Use of hybrid chitosan membranes and N1E-115 cells for promoting nerve

regeneration in an axonotmesis rat model. [Internet]. Biomaterials. 2008 Nov;

29(33):4409-19 DOI: 10.1016/j.biomaterials.2008.07.043.



1 Introduction

Peripheral nerve injuries are a frequent pathology in today’s society (1). Despite recent

progress in peripheral nerve trauma management, recovery of functional parameters is

usually far from normal, even for the most skilled surgeons, and thus much attention is

being paid to nerve regeneration research (2; 3).

The experimental model based on the induction of a crush injury (axonotmesis) in the

rat sciatic nerve provides a very realistic testing bench for lesions involving

plurifascicular mixed nerves with axons of different size and type competing to reach

and re-innervate distal targets (4; 5). After a lesion of axonotmesis, the distal nerve

fragment undergoes a process named Wallerian degeneration, which leads to the

degradation of axons and myelin sheaths and creates a favourable environment for

nerve regeneration (6-9). Both macrophages and Schwann cells are locally recruited to

eliminate axonal and myelin fragments. While distal stump degenerates, the proximal

stump initiates the regeneration process - the axonal ends elongate in order to reach

the distal stump and Schwann cells differentiate and multiply, being responsible for the

ensheathing and myelination of the newly sprouted axons (6; 7; 9-12). This type of

injury is thus appropriate to investigate the cellular and molecular mechanisms of

peripheral nerve regeneration, to assess the role of different factors in the regeneration

process (13) and to perform preliminary in vivo testing of biomaterials that will be useful

in tube-guide fabrication for more serious injuries of the peripheral nerve, such as

neurotmesis.

Autologous nerve grafting is the gold standard to reconstruct a large defect in a

peripheral nerve, but with some important disadvantages, such as availability of the

donor site and complications related to its sacrifice, inadequate recovery of function

and aberrant regeneration (14-21). Nowadays, the use of entubulation has attempted

to overcome these problems. A cylinder-shaped tube is placed between the nerve

ends, not only allowing orientation of growing nerve fibres, but also enabling the

incorporation of substances, either molecules or cells, that enhance nerve regeneration

(16; 21; 22). Among the various materials than can be used in the composition of the

tube guides, biodegradable substances offer two important advantages: one surgical

step is saved, as they avoid the need to be removed as required for autologus tissue

transplantation and it is possible to modulate the time of degradation according to the

axonal regeneration time diminishing inflammation on the lesion site. Thus, a major

challenge in tissue engineering is to create adequate scaffolds that are capable of

replace the autografts techniques. As far as peripheral nerve regeneration is

Chapter 3 – Use of hybrid chitosan membranes and N1E-115 cells for promoting nerve regeneration in an

axonotmesis rat model 41


concerned, a wide range of substances have been developed to meet this purpose (16;

21; 25; 26).

There are many properties required for desirable nerve guided conduit. They include

permeability that prevents fibrous scar tissue invasion but allow nutrient and oxygen

supply, revascularization to improve nutrient and oxygen supply, mechanical strengths

to maintain a stable support structure for the nerve regeneration, immunological

inertness with surrounding tissues, biodegradability to prevent chronic inflammatory

response and pain by nerve compression, easy regulation of conduit diameter and wall

thickness, and surgical amenability (26). The degradation rate of these biomaterials

should be related to the axonal regeneration time. Among the various substances

proposed for the fashioning of nerve conduits, chitosan has recently attracted particular

attention because of its biocompatibility, biodegradability, low toxicity, low cost,

enhancement of wound-healing and antibacterial effects(27). In addition, the potential

usefulness of chitosan in nerve regeneration have been demonstrated both in vitro and

in vivo (28-34). Chitosan is a partially deacetylated polymer of acetyl glucosamine

obtained after the alkaline deacetylation of chitin (35). Chitosan matrices have been

shown to have low mechanical strength under physiological conditions and to be

unable to maintain a predefined shape for transplantation, which has limited their use

as nerve guidance conduits in clinical applications. The improvement of their

mechanical properties can be achieved by modifying chitosan with a silane agent. γ-

glycidoxypropyltrimethoxysilane (GPTMS) is one of the silane-coupling agents, which

has epoxy and methoxysilane groups. The epoxy group reacts with the amino groups

of chitosan molecules, while the methoxysilane groups are hydrolyzed and form silanol

groups, and the silanol groups are subjected to the construction of a siloxane network

due to the condensation. Thus, the mechanical strength of chitosan can be improved

by the crosslinking between chitosan and GPTMS and siloxane network. Chitosan and

chitosan-based materials have been proven to promote adhesion, survival, and neurite

outgrowth of neural cells (36; 37).

Together with scaffolds, neurotrophic factors have also been the target of intensive

research - their role in nerve regeneration and the way they influence neural

development, survival, outgrowth, and branching (20). Among neurotrophic factors,

neurotrophins have been heavily investigated in nerve regeneration studies. They

include the nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF),

neurotrophin-3 (NT-3), and neurotrophin-4/5 (NT-4/5)(38). Neurotrophic factors

promote a variety of neural responses, including survival and outgrowth of the motor

and sensory nerve fibers, and spinal cord regeneration (22; 39). However, in vivo



responses to neurotrophic factors can vary due to the method of their delivering.

Therefore, the development and use of controlled delivery devices are required for the

study of complex systems. N1E-115 cell line that undergoes neuronal differentiation in

response to either dimethylsulfoxide (DMSO), adenosine 3’5’-cyclic monophosphate

(cAMP) or serum withdrawal is an important cellular system to locally produce and

deliver neurotrophic factors (40; 41).

Based on this premises, the aim of the study was to bring together two of the more

promising recent trends in nerve regeneration research: 1) local enwrapping of the

lesion site of axonotmesis by means of hybrid chitosan membranes; 2) application of a

cell delivery system to improve local secretion of neurotrophic factors.

First, types I, II and III chitosan membranes were screened by an in vitro assay. Then,

membranes were evaluated in vivo to assess their biocompatibility and their effects on

nerve fiber regeneration and nerve recovery in a standardized rat sciatic nerve crush

injury model (42; 43).




2 Materials and methods

2.1 Preparation of chitosan membranes

Chitosan (high molecular weight, Aldrich®, USA) was dissolved in 0.25M acetic acid

aqueous solution to a concentration of 2% (w/v). To obtain type II and type III

membranes, GPTMS (Aldrich®, USA) was also added to the chitosan solution and

stirred at room temperature for 1h. The solutions for type I and II chitosan membranes

were then poured into polypropylene containers with cover, and aged at 60°C for 2

days. The drying process for type III chitosan membrane was significantly different: the

solutions were frozen for 24h at -20°C and then transferred to the freeze-dryer, where

they were left 12h to complete dryness. The chitosan membranes (type I, II and III)

were soaked in 0.25N sodium hydroxide aqueous solution to neutralize remaining

acetic acid, washed well with distilled water, and dried again at 60°C for 2 days (type I

and II) or freeze dried (type III). All membranes were sterilized with ethylene oxide gas,

considered by some authors the most suitable method of sterilization for chitosan

membranes (44). Prior to their use in vivo, membranes were kept during 1 week at

room temperature in order to clear any ethylene oxide gas remnants.

Scanning electron microscope analysis

The surface morphology of the chitosan membranes was observed under a scanning

electron microscope (SEM; JEOL JSM 6301F) equipped with x-ray energy dispersive

spectroscopy (EDX) microanalysis capability (Voyager XRMA System, Noran

Instruments).

Cell culture, intracellular calcium concentration ([Ca2+]i) measurements and cell

adherence assays

N1E-115 is a clone of cells derived from mouse neuroblastoma C-1300 (45) and

retains numerous biochemical, physiological, and morphological properties of

differentiated neuronal cells in culture (46). N1E-115 neuronal cells were cultured in

poly-l-lisine coated Petri dishes (around 2 x 106 cells) on 2 x 2 cm chitosan fragments

(chitosan type I, type II and type III) at 37ºC, 5% CO2 in a humidified atmosphere

(Nuaire). Maintenance medium was 89.8% Dulbecco’s Modified Eagle’s Medium

(DMEM) supplemented with glutamine (GlutaMAX; Gibco), 10% fetal bovine serum

(FBS, Sigma), 0.1% penicillin/streptomycin (100000 U/ml penicillin, 10 mg/ml



streptomycin; Sigma) and with 0.1% β-amphoterrycin (250 μg/ml, Sigma). The culture

medium was changed every 48 hours and the cells were observed daily in an inverted

microscope. Before surgery, once N1E-115 cell culture reached approximately 80%

confluence, cells were supplied with differentiation medium containing DMSO. The

differentiation medium was composed by 95.8% Dulbecco’s Modified Eagle’s Medium

(DMEM) supplemented with glutamine (GlutaMAX; Gibco), 2.5% FBS, 0.1%

penicillin/streptomycin (100000 U/ml penicillin, 10 mg/ml streptomycin; Sigma), 0.1% β-

amphoterrycin (250 μg/ml, Sigma), and 1.5% DMSO (Sigma). Cell culture viability was

assessed by measuring intracellular free calcium concentration ([Ca2+]i). The [Ca2+]i

was measured in Fura-2-AM-loaded cells, through dual wavelength spectrofluorometry

as previously described (24). [Ca2+]i was determined in N1E-115 cell culture before

differentiation and 48 hours after the onset of DMSO-induced differentiation.

In vivo assays

All procedures were performed with the approval of the Veterinary Authorities of

Portugal in accordance with the European Communities Council Directive of November

1986 (86/609/EEC).

Figure 13 - The picture shows the dorsal incisions made on the dorsal

area of the four Wistar rats, to test the biocompatibility of the chitosan

membranes: 1 (left cranial incision), type I chitosan membrane; 2 (mid-

right incision), type II chitosan membrane; 3 (left caudal incision), type III

chitosan membrane




Biocompatibility assay

Prior to their use on crushed sciatic nerves, the three types of chitosan membranes

were tested in vivo to assess their biocompatibility: 4 adult female Wistar rats were

used. On each one, under general anaesthesia, 3 longitudinal dorsal incisions, 3 cm-

long, were made and 2 x 2 cm fragments were implanted (Figure 13). Animals were

sacrificed on weeks one, two, four and eight. The membrane remnants were collected

together with skin and subcutaneous tissues and were fixed in a 10% formaldehyde

solution for later histological analysis. Throughout the 8-week follow-up time, all

animals remained healthy, and none developed local or systemic signs of infection

and/or inflammation.

Nerve regeneration assay

In vivo nerve regeneration assay was carried out in types II and III chitosan

membranes only because of the higher elasticity which proved to facilitate surgery. A

total of 36 adult female Wistar rats (Charles River Laboratories, Barcelona, Spain)

weighing approximately 250g at the start of the experiment were used. The animals

were divided by six experimental groups of six animals each. Animals were housed two

animals per cage (Makrolon type 4, Tecniplast, VA, Italy), in a temperature and

humidity controlled room with 12-12h light / dark cycles, and were allowed normal cage

activities under standard laboratory conditions. The animals were fed with standard

chow and water ad libitum. Adequate measures were taken to minimize pain and

discomfort taking into account human endpoints for animal suffering and distress.

Animals were housed for 2 weeks before entering the experiment. All procedures were

performed with the approval from the Veterinary Authorities of Portugal, and in

accordance with the European Communities Council Directive of 24 November 1986

(86/609/EEC). The experimental groups were set according to treatment after nerve

sciatic crush injury. Therefore, in one group the animals recovered from the sciatic

crush injury without any other intervention (Crush). In other two groups, the crushed

sciatic nerve was encircled by a type II chitosan membrane either alone (ChitosanII) or

covered with a monolayer of N1E-115 cells, differentiated in vitro (ChitosanIICell). In

the remaining two groups, type III chitosan was used alone (ChitosanIII) or covered by

N1E-115 cells (ChitosanIIICell). Finally, an additional group of unoperated animals was

used as control for nerve histological analysis. The standardized crush injury was

carried out with the animals placed prone under sterile conditions and the skin from the



clipped lateral right thigh scrubbed in a routine fashion with antiseptic solution. Under

deep anaesthesia [ketamine (Imalgene 1000®) 9 mg/100 g; xylazine (Rompun®), 1.25

mg/100 g, atropine 0.025 mg/100 g body weight, IP], the right sciatic nerve was

exposed unilaterally through a skin incision extending from the greater trochanter to the

mid-thigh followed by a muscle splitting incision. After nerve mobilisation, a non-

serrated clamp (Institute of Industrial Electronic and Material Sciences, University of

Technology, Vienna, Austria) exerting a constant force of 54 N, was used for a period

of 30 seconds to create a 3-mm-long crush injury, 10 mm above the bifurcation into

tibial and common peroneal nerves (16; 20). The starting diameter of the sciatic nerve

was about 1 mm, flattening during the crush to 2 mm, giving a final pressure of p9

MPa. The nerves were kept moist with 37ºC sterile saline solution throughout the

surgical intervention. Muscle and skin were then closed with 4/0 resorbable sutures.

The surgical procedure was performed with the aid of an M-650 operating microscope

(Leica Microsystems, Wetzlar, Germany). To prevent autotomy, a deterrent substance

was applied to rats’ right foot (21; 25). The animals were intensively examined for signs

of autotomy and contracture and none presented severe wounds (absence of a part of

the foot or severe infection) or contractures during the study.

2.2 Functional Assessment of Reinnervation

Motor performance and nociceptive function

All animals were tested preoperatively (week 0), and every week until week 8 and then

every two weeks until the end of the 12-week follow-up time. Animals were gently

handled, and tested in a quiet environment to minimize stress levels. Motor

performance and nociceptive function were evaluated by measuring extensor postural

thrust (EPT) and withdrawal reflex latency (WRL), respectively. For EPT test, the entire

body of the rat, except the hindlimbs, was wrapped in a surgical towel and supported

by the thorax. The affected hindlimb was then lowered towards the platform of a digital

balance (model PLS 510-3, Kern & Sohn GmbH, Kern, Germany) to elicit the EPT. As

the animal was lowered over the platform, it extended the hindlimb, anticipating the

contact made by the distal metatarsus and digits. The force in grams applied to the

digital platform balance was recorded (digital scale range 0-500 g). The same

procedure was applied to the contra-lateral, unaffected limb. The affected and normal

limbs were tested 3 times, with an interval of 2 minutes between consecutive tests, and

the three values were averaged to obtain a final result. The normal (unaffected limb)

EPT (NEPT) and experimental EPT (EEPT) values were incorporated into an equation




Equation 3) to derive the percentage of functional deficit, as described in the literature

(27):

% Motor deficit = [(NEPT – EEPT)/NEPT] x 100

Equation 3 – EPT formula

The nociceptive withdrawal reflex (WRL) was adapted from the hotplate test as

described by Masters et al. (47). The rat was wrapped in a surgical towel above its

waist and then positioned to stand with the affected hindpaw on a hotplate at 56ºC

(model 35-D; IITC Life Science Instruments, Woodland Hill, CA). WRL is defined as the

time elapsed from the onset of hotplate contact to withdrawal of the hindpaw and

measured with a stopwatch. Normal rats withdraw their paws from the hotplate within 4

seconds or less (28). The affected limbs were tested three times, with an interval of 2

minutes between consecutive tests to prevent sensitization, and the three latency times

were averaged to obtain a final result (22). The cut off time for heat stimulation was set

at 12 seconds to avoid skin damage to the foot (28).

Sciatic Functional Index (SFI) and Static Sciatic Index (SSI)

For SFI, animals were tested in a confined walkway measuring 42-cm-long and 8.2-cm-

wide, with a dark shelter at the end. A white paper was placed on the floor of the rat

walking corridor. The hindpaws of the rats were pressed down onto a finger paint-

soaked sponge, and they were then allowed to walk down the corridor leaving its hind

footprints on the paper. Often, several walks were required to obtain clear print marks

of both feet. Prior to any surgical procedure, all rats were trained to walk in the corridor,

and a baseline walking track was recorded. Subsequently, walking tracks were

recorded every week until the week-8 postoperatively and then on weeks 10 and 12.

Several measurements were taken from the footprints: (I) distance from the heel to the

third toe, the print length (PL); (II) distance from the first to the fifth toe, the toe spread

(TS); and (III) distance from the second to the fourth toe, the intermediary toe spread

(ITS). For both dynamic (SFI) and static assessment (SSI), all measurements were

taken from the experimental (E) and normal (N) sides. Prints for measurements were

chosen at the time of walking based on clarity and completeness at a point when the

rat was walking briskly. The mean distances of three measurements were used to

calculate the following factors (dynamic and static):

Toe spread factor (TSF) = (ETS – NTS)/NTS

Intermediate toe spread factor (ITSF) = (EITS – NITS)/NITS

Print length factor (PLF) = (EPL – NPL)/NPL



Where the capital letters E and N indicate injured (experimental) and non-injured side

(normal), respectively.

SFI was calculated as described by Bain et al. (30) according to the following equation

Equation 4):

SFI = -38.3 (EPL – NPL)/NPL + 109.5(ETS-NTS)/NTS + 13.3(EIT-NIT)/NIT – 8.8 = (-

38.3 x PLF) + (109.5 x TSF) + (13.3 x ITSF) – 8.8

Equation 4- SFI formula

Static footprints were obtained at least during four occasional rest periods. For the

sciatic static index (SSI) only the parameters TS and ITS, were measured (31):

SSI = [(108.44 x TSF) + (31.85 x ITSF)] - 5.49

Equation 5 –SSI formula

For both SFI and SSI, an index score of 0 is considered normal and an index of -100

indicates total impairment. When no footprints were measurable, the index score of -

100 was given (32). Reproducible walking tracks could be measured from all rats. In

each walking track three footprints were analysed by a single observer, and the

average of the measurements was used in SFI calculations.

Kinematic analysis

Ankle kinematics and stance duration analysis were carried out prior to nerve injury

and on weeks one, four, eight and twelve of recovery. Animals walked on a perspex

track with length, width and height of respectively 120, 12, and 15 cm. In order to

ensure locomotion in a straight direction, the width of the apparatus was adjusted to the

size of the rats during the experiments, and a darkened cage was connected at the end

of the corridor to attract the animals. The rats’ gait was video recorded at a rate of 100

Hz images per second (JVC GR-DVL9800, New Jersey, USA). The camera was

positioned at the track half length where gait velocity was steady, and 1 m distant from

the track obtaining a visualization field of 14 cm wide. Only walking trials with stance

phases lasting between 150 and 400 ms were considered for analysis, since this

corresponds to the normal walking velocity of the rat (20–60 cm/s) (33; 34; 48). The

video images were stored in a computer hard disk for latter analysis using an


Diego, USA). 2-D biomechanical analysis (sagittal plan) was carried out applying a two-

segment model of the ankle joint, adopted from the model firstly developed by Varejão




et al. (48). Skin landmarks were tattooed at 3 points: the proximal edge of the tibia, the

lateral malleolus and, the fifth metatarsal head. The rats’ ankle angle was determined

using the scalar product between a vector representing the foot and a vector

representing the lower leg. Four complete walking cycles were analysed per rat. With

this model, positive and negative values of position of the ankle joint indicate

dorsiflexion and plantarflexion, respectively. For each stance phase the following time

points were identified: initial contact (IC), opposite toe-off (OT), heel-rise (HR) and toe-

off (TO) (34; 48), and were time normalized for 100% of the stance phase. The

normalized temporal parameters were averaged over all recorded trials. Angular

ankle’s velocity was also determined (negative values correspond to dorsiflexion).

2.3 Design-based quantitative morphology and electron

microscopy

After the 12-week follow-up time, rats were anaesthetised and a 10-mm-long segment

of the sciatic nerve that included the injured portion was collected, fixed, and prepared

for morphological analysis and histomorphometry of myelinated nerve fibers. A 10-mm

segment of uninjured sciatic nerve was also withdrawn from the 6 control animals.

Immediately after collecting the nerve, rats were euthanized through an intracardiac

injection of 5% sodium pentobarbital (Eutasil®). Sciatic nerve samples were immersed

immediately in a fixation solution, containing 2.5% purified glutaraldehyde and 0.5%

saccarose in 0.1M Sorensen phosphate buffer for 6-8 hours. Specimens were then

washed in a solution containing 1.5% saccarose in 0.1M Sorensen phosphate buffer,

post-fixed in 2% osmium tetroxide, dehydrated and embedded in Glauerts' embedding

mixture of resins consisting in equal parts of Araldite M and the Araldite Härter, HY 964

(Merck, Darmstad, Germany), to which was added 1-2% of the accelerator 964, DY

064 (Merck, Darmstad, Germany). The plasticizer dibutyl phthalate was added in a

quantity of 0.5% [75]. Series of 2-µm thick semi-thin transverse sections were cut using

a Leica Ultracut UCT ultramicrotome (Leica Microsystems, Wetzlar, Germany) and

stained by Toluidine blue for 2-3 minutes for high resolution light microscopy

examination. In each nerve, histomorphometry was conducted using a DM4000B

microscope equipped with a DFC320 digital camera and an IM50 image manager

system (Leica Microsystems, Wetzlar, Germany). This system reproduced microscopic

images (obtained through a 100x oil-immersion objective) on the computer monitor at a

magnification adjusted by a digital zoom. The final magnification was 6600x enabling



accurate identification and morphometry analysis of myelinated nerve fibers. One semi-

thin section from each nerve was randomly selected and used for the morpho-

quantitative analysis. The total cross-sectional area of the nerve was measured and

sampling fields were then randomly selected using a protocol previously described

(37). Possible "edge effects" were compensated by employing a two-dimensional

dissector procedure which is based on sampling the "tops" of fibers (37). Mean fiber

density was calculated by dividing the total number of nerve fibers within the sampling

field by its area (N/mm2). Total fibers number (N) was then estimated by multiplying the

mean fiber density by the total cross-sectional area of the whole nerve cross section

assuming a uniform distribution of nerve fibers across the entire section. Fiber and

axon area were measured and the circle-fitting diameter of fiber (D) and axon (d) were

calculated. These data were used to calculate myelin thickness [(D-d)/2], myelin

thickness/axon diameter ratio [(D-d)/2d], and fiber/axon diameter ratio (D/d). The

precision of the histomorphometry methods was evaluated by calculating the coefficient

of error (CE). Regarding quantitative estimates of fiber number, the CE(n) was

obtained as follows in Equation 6 (37):

'

1)(

QnCE

Equation 6 - quantitative estimates of fiber number

Where Q' is the number of counted fibers in all dissectors.

For size estimates, the coefficient of error was estimated as follows in Equation 7 (36):

Mean

SEMzCE )(

Equation 7 - coefficient of error

Where SEM = standard error of the mean.

The sampling scheme was designed in order to keep the CE below 0.10, which

assures enough accuracy for neuromorphological studies (49).

For transmission electron microscopy, ultra-thin sections were cut by means of the

same ultramicrotome and stained with saturated aqueous solution of uranyl acetate

and lead citrate. Ultra-thin sections were analyzed using a JEM-1010 transmission

electron microscope (JEOL, Tokyo, Japan) equipped with a Mega-View-III digital

camera and a Soft-Imaging-System (SIS, Münster, Germany) for the computerized

acquisition of the images.




3 Statistics

Two-way mixed factorial ANOVA was used to test for the effect of time (within subjects

effect) and group (between subjects effect). Sphericity was evaluated by the Mauchly’s

test and when this assumption was not satisfied, the degrees of freedom were

corrected by using the more conservative Greenhouse-Geiser’s epsilon. Differences

between pre-surgery results and those obtained throughout the 12-week recovery

period were systematically assessed by applying planned contrasts (General Linear

Model, simple contrasts). The effect of the chitosan membrane alone or associated to

N1E-115 differentiated cells was then evaluated through two way mixed factorial

ANOVA. For histomorphometry, statistical comparisons of quantitative data were

subjected to one-way ANOVA test. MANOVA analysis was employed to assess

differences in functional recovery between the experimental groups. Statistical

significance was established as p<0.05. All statistical procedures were performed by

using the statistical package SPSS (version 14.0, SPSS, Inc) except histomorphometry

data that was analysed using the software “Statistica per discipline bio-mediche”

(McGraw-Hill, Milan, Italy). All data in this study is presented as mean standard error

of the mean (SEM).



4 Results

4.1 In vitro results and SEM analysis of chitosan membranes

Results obtained from epifluorescence technique are referred to measurements from

non-differentiated N1E-115 cells and after 48h of differentiation in the presence of 1.5%

DMSO. The mean value of [Ca2+]i in non-differentiated N1E-115 cells (N = number of

cells submitted to [Ca2+]i measurement) was 39.9 ± 3.4 nM (N = 15), 35.9 ± 3.2 nM (N =

15) and 40.2 ± 2.9 nM (N = 15), for cultures over chitosan membranes, type I,II and III,

respectively. Values of [Ca2+]i for N1E-115 cells after 48h of differentiation in the

presence of 1.5% DMSO were 42.9 ± 5.1 nM (N = 15), 44.3 ± 4.8 nM (N = 15) and 41.6

± 4.3 nM (N = 15), for cultures over chitosan membranes, type I,II and III, respectively.

All these values are not statistically different for p<0.05, and correspond to [Ca2+]i from

cells that did not begin the apoptosis process besides the evident neural differentiation.

According to this fact, it is reasonable to conclude that chitosan membranes, previously

presented as type I, II and III, were a viable substrate for N1E-115 neuronal cell line

adhesion, multiplication and differentiation.

Figure 14 shows the SEM microstructure of type II Figure 14A) and type III (Figure

14B) chitosan membranes, respectively. The drying techniques employed to prepare

the membranes were freeze-drying (type III) and the conventional thermal drying (type I

and type II), which led to extremely dissimilar microstructural and mechanical

properties. In this study, the chitosan membranes type III have about 110 µm pores

and about 90% of porosity.

Figure 14 - Scanning electron microscopy microstructure of chitosan membranes. (A) Type II chitosan

membrane. (B) Type III chitosan membrane, showing a more porous microstructure, when compared to Type

II chitosan membrane




4.2 In vivo biocompatibility results

Despite the same composition, type II and type III membranes presented a distinct

behavior: the latter elicited an exuberant cellular infiltrate composed in a large extent by

multinucleated giant cells and some mast cells, whereas type II chitosan elicited a mild

fibrous capsule and a discrete inflammatory reaction. Type III chitosan membranes

underwent a completely different aging process, the so called freeze drying. This

lyophilisation procedure resulted in highly porous membranes: after freezing and

lyophilisation, the spaces formerly occupied by the solvent were left emptied so these

porous membranes presented a superior surface/volume ratio, when compared to the

type I and type II chitosan membranes. As surface/volume ratio increases, there is a

higher contact surface with the host’s immune system, which could explain the

resulting exuberant cellular component. In fact, this study demonstrated that the three

chitosan membranes tested in vitro and in vivo were biocompatible and therefore an

important scaffold for the reconstruction of peripheral nerve, after axonotmesis or

neurotmesis injury, associated or not to neurotrophic factors cellular producing

systems. Types II and III chitosan membranes showed higher elasticity in comparison

to type I chitosan membrane, a feature which would facilitate the surgical manipulation.

Therefore, we decided to carry out in vivo testing with chitosan types II and III only.

4.3 Functional Assessment of Reinnervation

Immediately after the acute compression injury, the crushed areas of all sciatic nerves

were flattened but nerve continuity was preserved. Complete flaccid paralysis of the

operative foot was observed following crush injury. All rats survived, with no wound

infection or automutilation.

Withdrawal reflex latency (WRL)

The results obtained, performing Withdrawal Reflex Latency (WRL) test to evaluate the

nociceptive function are shown in Figure 15. Two way ANOVA shows that the WRL

were significantly affected after the sciatic nerve crush [F(10,250)=82,884; p<0,000].

Contrast analysis indicates that the WRL values were affected up to week 8. Beyond

this time point, WRL values, the experimental groups pooled together, were similar to

those preoperatively. A significant effect of group was found for WRL data

[F(4,25)=12,018; p<0,000], and post hoc analysis showed that differences were

significant between ChitosanII group from one side and ChitosanIICell, ChitosanIII and



ChitosanIIICell groups on the other side (p<0,05). These results are suggestive of a

slower rate of recovery in WRL with the use of type II chitosan.

Figure 15 - WRL results

Motor Deficit (EPT)

The EPT values obtained during the healing period of 12 weeks are represented in

Figure 16. EPT values were significantly affected by the crush injury [F(10,250)=1497,366;

p<0,000] and contrast analysis shows that at week 12, EPT had still not recovered to

baseline values. There were significant differences in EPT values between the

experimental groups [F(4,25)=12,018; p<0,000] and post hoc analysis shows that the

EPT values from the Crush group were significantly different from those of the

ChitosanIICell and ChitosanIIICell (p<0,05), suggesting a negative effect of the N1E

115-differentiated cells on motor recovery after sciatic crush injury.

Figure 16 - EPT results




Sciatic functional index (SFI) and Static Sciatic Index (SSI)

SFI and SSI results are depicted in Figure 17 and Figure 18, respectively. Two way

ANOVA shows that SFI values changed significantly after the crush injury

[F(10,250)=488,931; p<0,000]. Contrast analysis shows that SFI values were different

from baseline until week 7 of recovery. A significant effect of group was observed for

SFI values [F(4,25)=144,125; p=0,01] with post hoc analysis indicating that values from

ChitosanIICell group were different from those of the ChitosanII and ChitosanIII groups.

Two way ANOVA analysis on SSI data was similar to that of the just reported SFI

results. However, post hoc analysis shows that SSI results from the ChitosanIIICell

group were different from those of the Crush, ChitosanII and ChitosanIICell groups.

Figure 17 - SFI results



Figure 18 - SSI results

Ankle joint Kinematics

Video recordings of the rats’ gait were undertaken at week 0 (pre-operatively) and at

post-surgery at weeks 0, 2, 4, 8 and 12, so the amount of time points for statistical

analysis was reduced to only five (see Methods). Figure 19 and Figure 20 represent

the values of ankle joint angle and the values of ankle joint velocity at IC, OT, HR and

TO, respectively, obtained by kinematic analysis during the healing period of 12 weeks.




Figure 19 - Curves representing ankle joint angle in the sagittal plane during the stance phase of

walking. Note the absence of plantarflexion at push off in all sciatic crushed groups at week 2

post-injury. The normal ankle motion pattern is progressively regained during the 12-weeks

recovery period. The mean of each group is plotted: (solid line) Control Group; () Axonotmesis

Group (Crush); (dashed line) Axonotmesis involved with chitosan II membrane (ChitosanII); ()

Axonotmesis involved with chitosan III membrane (ChitosanIII); (x) Axonotemesis involved with

chitosan type II membrane covered with the cellular system (ChitosanIICell); () Axonotemesis

involved with chitosan type III membrane covered with the cellular system (ChitosanIIICell).



Initial Contact (IC)

The ankle joint angle at IC changed significantly after the sciatic crush [F(4, 100)=28,705;

p<0,000]. Such changes were transient and by week 4 ankle joint angle at IC was

indistinguishable from baseline, as indicated by contrast analysis. Two way ANOVA

revealed group differences [F(4,25)=10,671; p<0,000] and post hoc analysis shows that

ChitosanII group was different from all the other four groups.

Opposite toe off (OT)

Ankle joint angle at OT was significantly affected after the crush injury [F(4,100)=13,464;

p<0,000] but by week 4 of recovery joint position at this point of stance had recovered

to normal values. Two way ANOVA revealed a significant group effect [F(4,25)=4,150;

p=0,01] and post hoc analysis showed that ankle joint angles at OT in the ChitosanII

group throughout the study were significantly different from Crush and ChitosanIII

groups (Figure 19).

Ankle joint velocity at OT was significantly affected after the crush injury [F(4,100)=8,582;

p<0,000] and returned to preoperative values at week 4 (p<0,05). There were

significant differences between the groups for ankle joint velocity at OT [F(4,25)=3,164;

p<0,031] with post hoc analysis showing significant differences between ChitosanII and

ChitosanIICell groups (p<0,05) (Figure 20).

Heel Raise (HR)

Ankle joint angle at HR was significantly altered after the sciatic nerve injury

[F(4,100)=34,151; p<0,000] until week 12 (p<0,05). A significant effect of group was found

for this variable [F(4,25)=3,456; p<0,022]. Post hoc tests found significant differences in

ankle joint angle at HR between ChitosanII and the Crush and ChitosanIIICell groups

(Figure 19).

Instantaneous ankle’s joint velocity at HR was significantly affected after axonotmesis

[F(4,100)=42,384; p<0,000] with contrast analysis showing that differences in ankle’s joint

velocity were significantly different from normal values until week 4 of recovery. A

significant effect of group was registered [F(4,25)=3,766; p<0,016], and post hoc tests

showed significant differences between ChitosanII and ChistosanIICell groups (Figure

20).




Figure 20 - Curves represent ankle joint angular velocity in the sagittal plane during the stance

phase of walking. Note the loss of the normal ankle joint angular velocity biphasic response in all

sciatic crushed groups at week 2 post-injury. The normal ankle velocity pattern is progressively

regained during the 12-weeks recovery period. The mean of each group is plotted: (solid line)

Control Group; ()Axonotmesis Group (Crush); (dashed line) Axonotmesis involved with chitosan

II membrane (ChitosanII); () Axonotmesis involved with chitosan III membrane (ChitosanIII); (x)

Axonotemesis involved with chitosan type II membrane covered with the cellular system

(ChitosanIICell); ()Axonotemesis involved with chitosan type III membrane covered with the

cellular system (ChitosanIIICell).



Toe off (TO)

Ankle joint angle at TO was significantly altered after the sciatic nerve injury

[F(4,100)=181,588; p<0,000]. Contrast analysis showed that at week 12 ankle’s joint

angle mean value at TO was similar to that registered before the sciatic nerve injury. A

significant effect of group was found for this variable [F(4,25)=3,409; p<0,023]. Post hoc

tests found significant differences in ankle’s joint angle at TO between ChitosanII and

ChitosanIIICell groups (Figure 19). Ankle’s joint velocity at TO was significantly affected

after the crush injury [F(4,100)=103,331; p<0,000] with contrast analysis showing that

ankle’s joint velocity did not recovered completely within the 12-weeks follow up time. A

significant effect of group was registered [F(4,25)=9,689; p<0,000], and post hoc tests

showed significant differences between ChitosanIICell, from one hand and Crush,

ChitosanIII and ChistosanIIICell groups, on the other hand (Figure 20).

MANOVA analysis was employed to further compare the results of the walk kinematic

analysis of the experimental groups, using values of week 12 only. A significant effect

of the experimental group was observed (Pillai’s Trace 3,709; p<0,000). Multivariate

contrast analysis (K matrix) showed that ChitosanIII was the group differing from Crush

group in a less number of variables. This indicates that at the end of the recovery

period animals in ChitosanIII group presented a degree of functional recovery similar to

that registered in the Crush group.

4.4 Morphological and histomorphometrical analysis

Figure 21 shows the histological appearance of control normal sciatic nerve cross

sections (Figure 21A) in comparison to cross sections of the regenerated nerve fibers

with and without cell delivery (Figure 21B-F). As expected, regenerated nerve fibers in

all five experimental groups were organized in microfascicles and were smaller than

control nerve fibers. In the two experimental groups in which cell delivery was carried

out (Figure 21B,D,F), the presence of unusual cell profiles, interpretable as the

transplanted cells, were detectable at the periphery of the nerves. This was seldom

observed in inner parts of the nerve (Figure 21F) suggesting that only few transplanted

cells colonized the inner nerve interstice.

Electron microscopy confirms that a good regeneration pattern of both unmyelinated

and myelinated nerve fibers occurred in all experimental groups, irrespectively of the

enrichment with neural stem cells (Figure 23A,B and Figure 24A,B). The border zone of

the nerve still shows the presence of some chitosan debris integrated with the collagen

fibers of the epineurium (Figure 23C,D). Moreover, in the two experimental groups in

which cell delivery was carried out (Figure 24), it was possible to find some of




transplanted cells localized in the border zone of the regenerated nerve (Figure

24C,D).

Results of the design-based morphoquantitative analysis of regenerated myelinated

nerve fibers permitted to quantitatively compare the different experimental groups. As

expected, fiber density was significantly (p<0.05) higher in all experimental groups in

comparison to control sciatic nerves, while mean axon and fiber diameter and myelin

thickness were significantly (p<0.05) lower.



As regards the number of regenerated myelinated nerve fibers, this parameter was

found to be significantly higher (p<0.05) in comparison to controls in four out of the five

experimental groups only; in fact, ChitosanIII group showed a number of fibers not

significantly (p<0.05) different in comparison to normal control nerves.

Figure 21 - High resolution photomicrographs of nerve fibers from control and regenerated rat sciatic

nerves. (A) Control group; (B) Crush group; (C) ChitosanII group; (D) ChitosanIICell group; (E) ChitosanIII

group. (F) ChitosanIIICell group. Original magnification: 1000x




The peculiarity of the experimental group characterized by type III chitosan membrane

enwrapment, was also confirmed by statistical analysis of inter-group variability among

experimental groups which demonstrated that ChitosanIII group has a significantly

(p<0.05) higher mean fiber and axon diameter and myelin thickness in comparison with

all other experimental crush groups.

Figure 22 - High resolution photomicrographs comparing ChitosanIII (A, C, E) and

ChitosanIIICell (B, D, F) groups. The presence of transplanted cells is clearly identifiable at

the periphery of the nerve (D) and, seldom, also in the inner nerve areas (F). Original

magnification: A, B: = 200x; C-F=1000x



Figure 23 - Electron microscopy regenerated nerve fibers in ChitosanIII experimental group.

In panels A and B, the ultrastructural appearance of unmyelinated (A) and myelinated (B)

nerve fibers can be appreciated. In panels C and D, the presence of chitosan debris mingled

with epineurial connective tissue can be still detected. Original magnification: A = 40,000x; B

= 30,000x; C = 50,000x; D = 150,000x.

A B

C D




Figure 24 - Electron microscopy regenerated nerve fibers in ChitosanIIICell experimental group. In panels A

and B, the ultrastructural appearance of unmyelinated (A) and myelinated (B) nerve fibers can be

appreciated. In panels C and D, the presence of elongated transplanted cells mingled with chitosan debris

and epineurial connective tissue can be still detected. Original magnification: A = 40,000x; B = 25,000x; C =

10,000x; D = 15,000x.



5 Discussion

To find out effective strategies for promoting peripheral nerve regeneration is a

challenging issue in Regenerative Medicine (1-3; 19; 38; 50-52). The study of new

materials for promoting nerve regeneration is usually carried out in axonotmesis nerve

lesion models (20). Although nerve fibers always regenerate after experimental

axonotmesis, there are considerable differences in the extent and rate of recovery

between studies which probably reflect the different techniques used to induce the

injury (20; 21; 38). To cope with this limitation, in this study we used a standardized

clamping procedure that was described in details in previous works (20; 21; 49; 53), for

investigating the effects of enwrapping the axonotmesis site with type II and type III

chitosan membranes. In two experimental groups, chitosan membranes were covered

with in vitro differentiated N1E-115 cells before implanting them in vivo. In neuronal

cells the regulation of the intracellular free calcium concentration [Ca2+]i plays an

important role in physiological processes such as growth and differentiation, controlling

important cell functions like the release of neurotransmitters and the membrane’s

excitability. The mechanisms that control [Ca2+]i are of crucial importance for normal

homeostasis, and its deregulation has been associated to cellular changes and even

cell death, when [Ca2+]i reaches values above 105 nM (42). The measurements of

[Ca2+]i of the N1E-115 cells in vitro differentiated by the addition of DMSO indicated

that these cells did not begin the apoptosis process although the evident neural

differentiation supporting the hypothesis that chitosan membranes type I, II and III,

were a viable substrate for N1E-115 neuronal cell line adhesion, multiplication and

differentiation. The in vivo biocompatibility studies also corroborated that the three

types of chitosan membranes were biocompatible and therefore an important scaffold

for the reconstruction of peripheral nerve, after axonotmesis or neurotmesis injury,

associated or not to cellular systems producing neurotrophic factors. The higher

elasticity observed in type II and III chitosan membranes, when compared to type I, that

facilitates surgery, induced us to carry out the in vivo testing on the rat sciatic nerve

regeneration, with these GPTMS hybrid membranes only, abandoning type I

membranes. As a matter of fact, the GPTMS hybrid membranes provide a suitable

scaffolding environment for neural tissue engineering, making this material a potential

candidate for scaffolds in neural tissue engineering.

The in vivo study on crushed rat sciatic nerve experimental model showed that

morphological predictors of nerve regeneration (number of fibers, axon and fiber size

and myelin thickness) were significantly improved in ChitosanIII group in comparison to




all other experimental crush groups, thus suggesting that type III chitosan is a suitable

biomaterial to be used in reconstructive peripheral nerve surgery. Functional

assessment partially supported the morphometric results since results of the WRL,

EPT, SFI and SSI tests suggest that chitosan type II and the N1E-115 cells have a

reduced degree of functional recovery which is in line with results of histomorphometric

analysis.

Results of both morphological and functional predictors of nerve regeneration also

showed that enrichment of chitosan membranes with N1E-115 neural cells did not have

any positive effect on nerve regeneration in comparison to crush controls and, in case

of type III chitosan membrane, the presence of transplanted cells seemed to prevent

the positive effect of the membrane wrapping alone on nerve regeneration. These

results are in agreement with previous investigation that showed that N1E-115 cell

population does not have significant effect in promoting axon regeneration and, when

N1E-115 cells were cultured inside a PLGA scaffold used to bridge a nerve defect, they

can even exert negative effects on nerve fiber regeneration (21; 53). Thus N1E-115

cells did not prove to be a potential candidate therapeutic agent for treatment of nerve

injury and their utilization is just limited to research purposes, mainly as a basic

scientific step in the investigation of the potentiality for nerve regeneration promotion of

cell transplantation delivery systems that permit that transplanted cells are able to

secrete neurotrophic factors in the site of injury without being directly in contact with

regenerating axons.

Chitosan type II and chitosan type III were developed as a hybrid of chitosan by the

addition of GPTMS. Wettability of material surfaces is one of the key factors for protein

adsorption, cell attachment and migration (39). The addition of GPTMS improves the

wettability of chitosan surfaces (41), and therefore chitosan type II and chitosan type III

are expected to be more hydrophilic than the original chitosan (41). Chitosan type III

was developed to be more porous, with a larger surface to volume ratio but preserving

mechanical strength and the ability to adapt to different shapes. Significant differences

in water uptake between commonly used chitosan and our hybrid chitosan type III were

previously reported due to the difference in the ability of the matrix to hold water (13).

In fact, hybrid chitosan-based membranes may retain about two times as much

biological liquid fluid as chitosan (13). A synergetic effect of a more favourable porous

microstructure and physicochemical properties (more wettable and higher water uptake

level) of chitosan type III compared to chitosan type II, and the presence of silica ions

may be responsible for the good results obtained in this study for the former in the

nerve regeneration process.



The observation that significant improvement of axonal regeneration was obtained in

crushed sciatic nerves surrounded by chitosan type III membranes alone suggests that

this material may not just work as a simple mechanical scaffold but instead may work

as an inducer of nerve regeneration. The regenerative property of chitosan type III

might be explained by the action on Schwann cell proliferation and axon elongation and

myelination (9; 13). Yet, the expression of established myelin genes such as PMP22,

PO and MBP (40) may be influenced by the presence of silica ions which exert an

effect on several glycoprotein expression (13; 43).




6 Conclusions

Enwrapment of the site of a nerve crush lesion with a chitosan type III membrane

significantly improves nerve regeneration in the rat, whereas enrichment of chitosan

membranes with N1E-115 neural cells does have positive effects. Whereas translation

of animal studies to patients should always be dealt with caution, the results obtained

in this experimental study open interesting perspectives for the clinical employment of

hybrid chitosan membranes in peripheral nerve reconstruction.



REFERENCES

1. Schlosshauer B, Dreesmann L, Schaller H-E, Sinis N. Synthetic nerve guide

implants in humans: a comprehensive survey. [Internet]. Neurosurgery. 2006

Oct;59(4):740-7; discussion 747-8.

2. Samii M, Carvalho GA, Nikkhah G, Penkert G. Surgical reconstruction of the

musculocutaneous nerve in traumatic brachial plexus injuries. [Internet]. Journal

of neurosurgery. 1997 Dec;87(6):881-6.[

3. Höke A. Mechanisms of Disease: what factors limit the success of peripheral

nerve regeneration in humans? [Internet]. Nature clinical practice. Neurology.

2006 Aug;2(8):448-54.

4. Lee JM, Tos P, Raimondo S, Fornaro M, Papalia I, Geuna S, et al. Lack of

topographic specificity in nerve fiber regeneration of rat forelimb mixed nerves.

[Internet]. Neuroscience. 2007 Feb;144(3):985-90. 1

5. Tos P, Battiston B, Nicolino S, Raimondo S, Fornaro M, Lee JM, et al.

Comparison of fresh and predegenerated muscle-vein-combined guides for the

repair of rat median nerve. [Internet]. Microsurgery. 2007 Jan;27(1):48-55.

6. Chen G, Ushida T, Tateishi T. Preparation of poly(L-lactic acid) and poly(DL-

lactic-co-glycolic acid) foams by use of ice microparticulates. [Internet].

Biomaterials. 2001 Sep;22(18):2563-7.


artificial cells, and artificial organs. 1990 Jan;18(1):1-24.

8. Yamaguchi I, Itoh S, Suzuki M, Osaka A, Tanaka J. The chitosan prepared from

crab tendons: II. The chitosan/apatite composites and their application to nerve

regeneration. [Internet]. Biomaterials. 2003 Aug;24(19):3285-92.

9. Yuan Y, Zhang P, Yang Y, Wang X, Gu X. The interaction of Schwann cells with

chitosan membranes and fibers in vitro. [Internet]. Biomaterials. 2004

Aug;25(18):4273-8.

10. Ishikawa N, Suzuki Y, Ohta M, Cho H, Suzuki S, Dezawa M, et al. Peripheral

nerve regeneration through the space formed by a chitosan gel sponge.

[Internet]. Journal of biomedical materials research. Part A. 2007 Oct

;83(1):3340.

11. Wang W, Itoh S, Matsuda A, Ichinose S, Shinomiya K, Hata Y, et al. Influences

of mechanical properties and permeability on chitosan nano/microfiber mesh




tubes as a scaffold for nerve regeneration. [Internet]. Journal of biomechanical

materials research. Part A. 2008;84(2):557-66.


[Internet]. Advanced drug delivery reviews. 2004 Jun;56(10):1467-80.



membranes. [Internet]. Biomaterials. 2005 Feb;26(5):485-93.

14. Amano T, Richelson E, Nirenberg M. Neurotransmitter synthesis by

neuroblastoma clones (neuroblast differentiation-cell culture-choline

acetyltransferase-acetylcholinesterase-tyrosine hydroxylase-axons-dendrites).

[Internet]. Proceedings of the National Academy of Sciences of the United States

of America. 1972 Jan;69(1):258-63.

15. Marreco PR, Luz Moreira P da, Genari SC, Moraes AM. Effects of different

sterilization methods on the morphology, mechanical properties, and cytotoxicity

of chitosan membranes used as wound dressings. [Internet]. Journal of

biomedical materials research. Part B, Applied biomaterials. 2004

Nov;71(2):268-77. 9


clamp. [Internet]. Journal of reconstructive microsurgery. 2001 Oct;17(7):531-4.

17. Kimhi Y, Palfrey C, Spector I, Barak Y, Littauer UZ. Maturation of neuroblastoma

cells in the presence of dimethylsulfoxide. [Internet]. Proceedings of the National

Academy of Sciences of the United States of America. 1976 Feb;73(2):462-6.


regeneration. [Internet]. Annual review of biomedical engineering. 2003

Jan;5293-347.

19. Lundborg G, Dahlin L, Danielsen N, Zhao Q. Trophism, tropism, and specificity

in nerve regeneration. [Internet]. Journal of reconstructive microsurgery. 1994

Sep;10(5):345-54.



Neurological research. 2004 Mar;26(2):186-94.






methods. 2007 Jun;163(1):92-104.




2004 Nov;21(11):1652-70.

23. Tuttle JB, Richelson E. Phenytoin action on the excitable membrane of mouse

neuroblastoma. [Internet]. The Journal of pharmacology and experimental

therapeutics. 1979 Dec;211(3):632-7.

24. Fisher SK, Domask LM, Roland RM. Muscarinic receptor regulation of

cytoplasmic Ca2+ concentrations in human SK-N-SH neuroblastoma cells: Ca2+

requirements for phospholipase C activation. [Internet]. Molecular pharmacology.

1989 Feb;35(2):195-204.



Journal of neuroscience methods. 1991 Feb;36(2-3):263-5.



standardized crush injury in the rat. [Internet]. Microsurgery. 2008 Jan;28(6):458-

70.[cited 2011 Mar 15]


nerve transection. [Internet]. Experimental neurology. 2001 Mar;168(1):192-

5.[cited 2010 Dec 7]


after sciatic nerve blockade. [Internet]. Anesthesiology. 1997 Apr;86(4):957-65.




2003 Jan;23(4):346-53.

30. Bain JR, Mackinnon SE, Hunter DA. Functional evaluation of complete sciatic,

peroneal, and posterior tibial nerve lesions in the rat. [Internet]. Plastic and

reconstructive surgery. 1989 Jan;83(1):129-38.






neuroscience methods. 2000 Oct;102(2):109-16.

32. Dijkstra JR, Meek MF, Robinson PH, Gramsbergen A. Methods to evaluate

functional nerve recovery in adult rats: walking track analysis, video analysis and

the withdrawal reflex. [Internet]. Journal of neuroscience methods. 2000

Mar;96(2):89-96.

33. Clarke KA, Parker AJ. A quantitative study of normal locomotion in the rat.

[Internet]. Physiology & behavior. 1986 Jan;38(3):345-51.[



nerve. 2002 Nov;26(5):630-5.



observation. [Internet]. Microscopy research and technique. 2008 Jul;71(7):497-

502.

36. Schmitz C. Variation of fractionator estimates and its prediction. [Internet].

Anatomy and embryology. 1998 Nov;198(5):371-97.



Jan;24(1):72-6.



Nov;14(11):1116-22.

39. Hench L. Biomaterials: An Interfacial Approach (Biophysics and bioengineering

series) [Internet]. Academic Pr; 1982.

40. Kuhn G, Lie A, Wilms S, Müller HW. Coexpression of PMP22 gene with MBP

and P0 during de novo myelination and nerve repair. [Internet]. Glia. 1993

Aug;8(4):256-64.

41. Tateishi T, Chen G, Ushida T. Biodegradable porous scaffolds for tissue

engineering [Internet]. Journal of Artificial Organs. 2002 Jun;5(2):77-83.



42. Trump BF, Berezesky IK. Calcium-mediated cell injury and cell death. [Internet].

The FASEB journal: official publication of the Federation of American Societies

for Experimental Biology. 1995 Feb;9(2):219-28.

43. Pietak AM, Reid JW, Stott MJ, Sayer M. Silicon substitution in the calcium

phosphate bioceramics. [Internet]. Biomaterials. 2007 Oct ;28(28):4023-32.

44. Wenling C, Duohui J, Jiamou L, Yandao G, Nanming Z, Xiufang Z. Effects of the

degree of deacetylation on the physicochemical properties and Schwann cell

affinity of chitosan films. [Internet]. Journal of biomaterials applications. 2005

Oct;20(2):157-77.



peripheric nerve regeneration. [Internet]. Acta médica portuguesa.

2005;18(5):323-8.




medical materials and engineering. 2005 Jan;15(6):455-65.



biodegradable polymer matrix. [Internet]. Anesthesiology. 1993 Aug;79(2):340-6.



[Internet]. Muscle & nerve. 2003;27(6):706–714. l

49. Varejão ASP, Cabrita AM, Meek MF, Fornaro M, Geuna S, Giacobini-Robecchi

MG. Morphology of nerve fiber regeneration along a biodegradable poly (DLLA-


2003 Jan;23(4):338-45.

50. Geuna S, Papalia I, Tos P. End-to-side (terminolateral) nerve regeneration: a

challenge for neuroscientists coming from an intriguing nerve repair concept.

[Internet]. Brain research reviews. 2006 Sep;52(2):381-8.

51. Battiston B, Geuna S, Ferrero M, Tos P. Nerve repair by means of tubulization:

literature review and personal clinical experience comparing biological and

synthetic conduits for sensory nerve repair. [Internet]. Microsurgery. 2005

Jan;25(4):258-67.




52. Meek MF, Robinson PH, Stokroos I, Blaauw EH, Kors G, Dunnen WF den.

Electronmicroscopical evaluation of short-term nerve regeneration through a

thin-walled biodegradable poly(DLLA-epsilon-CL) nerve guide filled with modified

denatured muscle tissue. [Internet]. Biomaterials. 2001 May;22(10):1177-85.



reconstruction of the rat sciatic nerve [Internet]. Microsurgery. 2007;27(2):125–

137.

Chapter 4 - Effects of Collagen Membranes Enriched

with in Vitro-Differentiated N1E-115 Cells on Rat Sciatic

Nerve Regeneration after End-to-End Repair

Amado S, Rodrigues JM, Luís AL, Armada-da-Silva P a S, Vieira M, Gartner A, et al.

Effects of collagen membranes enriched with in vitro-differentiated N1E-115 cells on rat

sciatic nerve regeneration after end-to-end repair. [Internet]. Journal of

neuroengineering and rehabilitation. 2010 Jan;7:7.



5 Introduction

Nerve regeneration is a complex biological phenomenon. In the peripheral nervous

system, nerves can spontaneously regenerate without any treatment if nerve continuity

is maintained (axonotmesis) whereas more severe type of injuries must be surgically

treated by direct end-to-end surgical reconnection of the damaged nerve ends (1-3).

Unfortunately, the functional outcomes of nerve repair are in many cases unsatisfactory

(4-10) thus calling for research in order to reveal more effective strategies for improving

nerve regeneration. However, recent advances in neuroscience, cell culture, genetic

techniques, and biomaterials provide optimism for new treatments for nerve injuries (1;

9-20).

The use of materials of natural origin has several advantages in tissue engineering.

Natural materials are more likely to be biocompatible than artificial materials. Also, they

are less toxic and provide a good support to cell adhesion and migration due to the

presence of a variety of surface molecules. Drawbacks of natural materials include

potential difficulties in their isolation and controlled scale-up (6-10). In addition to the

use of intact natural tissues, a great deal of research has focused on the use of purified

natural extracellular matrix (ECM) molecules, which can be modified to serve as

appropriate scaffolding (16). ECM molecules, such as laminin, fibronectin and collagen

have also been shown to play a significant role in axonal development and

regeneration (8; 21; 22). For example, silicone tubes filled with laminin, fibronectin, and

collagen led to a better regeneration over a 10 mm rat sciatic nerve gap compared to

empty silicone controls (14). Collagen filaments have also been used to guide

regenerating axons across 20–30 mm defects in rats (23-26). Further studies have

shown that oriented fibers of collagen within gels, aligned using magnetic fields,

provide an improved template for neurite extension compared to randomly oriented

collagen fibers (22; 27). Finally, rates of regeneration comparable to those using a

nerve autograft have been achieved using collagen tubes containing a porous

collagen-glycosaminoglycan matrix (28-30). Nerve regeneration requires a complex

interplay between cells, ECM, and growth factors. The local presence of growth factors

plays an important role in controlling survival, migration, proliferation, and differentiation

of the various cell types involved in nerve regeneration (8-10; 31). Therefore, therapies

with relevant growth factors received increasing attention in recent years although

growth factor therapy is a difficult task because of the high biological activity (in pico- to

nanomolar range), pleiotrophic effects (acting on a variety of targets), and short

Chapter 4 - Effects of Collagen Membranes Enriched with in Vitro-Differentiated N1E-115 Cells on Rat Sciatic

Nerve Regeneration after End-to-End Repair 79


biological half-life (few minutes to hours) (32). Thus, growth factors should be

administered locally to achieve an adequate therapeutic effect with little adverse

reactions and the short biological half-life of growth factors demands for a delivery

system that slowly releases locally the molecules over a prolonged period of time.

Employment of biodegradable membranes enriched with a cellular system producing

neurotrophic factors has been suggested to be a rational approach for improving nerve

regeneration after neurotmesis (16).

The aim of this study was thus to verify if rat sciatic nerve regeneration after end-to-end

reconstruction can be improved by seeding in vitro differentiated N1E-115 neural cells

on a type III equine collagen membrane and enwrap the membrane around the lesion

site. The N1E-115 cell line has been established from a mouse neuroblastoma (33)

and have already been used with conflicting results as a cellular system to locally

produce and deliver neurotrophic factors (6-10). In vitro, the N1E-115 cells undergo

neuronal differentiation in response to dimethylsulfoxide (DMSO), adenosine 3’, 5’-

cyclic monophosphate (cAMP), or serum withdrawal (4-10). Upon induction of

differentiation, proliferation of N1E-115 cells ceases, extensive neurite outgrowth is

observed and the membranes become highly excitable (4-10). The interval period of 48

hours of differentiation was previously determined by measurement of the intracellular

calcium concentration (Ca2+i). At this time point, the N1E-115 cells present already

the morphological characteristics of neuronal cells but cell death due to increased

Ca2+i is not yet occurring as described elsewhere (4-10).



6 Methods

6.1 Cell culture

The N1E-115 cells, clones of cells derived from the mouse neuroblastoma C-130035

retain numerous biochemical, physiological, and morphological properties of neuronal

cells in culture (4-10). N1E-115 neuronal cells were cultured in Petri dishes (around 2 x

106 cells) over collagen type III membranes (Gentafleece®, Resorba Wundversorgung

GmbH + Co. KG, Baxter AG) at 37ºC, 5% CO2 in a humidified atmosphere with 90%

Dulbecco’s Modified Eagle’s Medium (DMEM; Gibco) supplemented with 10% fetal

bovine serum (FBS, Gibco), 100 U/ml penicillin, and 100 µg/ml streptomycin (Gibco).

The culture medium was changed every 48 hours and the Petri dishes were observed

daily. The cells were passed or were supplied with differentiating medium containing

1.5% of DMSO once they reached approximately 80% confluence, mostly 48 hours

after plating (and before the rats’ surgery). The differentiating medium was composed

of 96% DMEM supplemented with 2.5 % of FBS, 100 U/ml penicillin, 100 µg/ml

streptomycin and 1.5% of DMSO (6-10).

6.2 Surgical procedure

Adult male Sasco Sprague Dawley rats (Charles River Laboratories, Barcelona, Spain)

weighing 300-350 g, were randomly divided in 3 groups of 6 or 7 animals each. All

animals were housed in a temperature and humidity controlled room with 12-12 hours

light / dark cycles, two animals per cage (Makrolon type 4, Tecniplast, VA, Italy), and

were allowed normal cage activities under standard laboratory conditions. The animals

were fed with standard chow and water ad libitum. Adequate measures were taken to

minimize pain and discomfort taking in account human endpoints for animal suffering

and distress. Animals were housed for two weeks before entering the experiment. For

surgery, rats were placed prone under sterile conditions and the skin from the clipped

lateral right thigh scrubbed in a routine fashion with antiseptic solution. The surgeries

were performed under an M-650 operating microscope (Leica Microsystems, Wetzlar,

Germany). Under deep anaesthesia (ketamine 90 mg/Kg; xylazine 12.5 mg/Kg,

atropine 0.25 mg/Kg i.m.), the right sciatic nerve was exposed through a skin incision

extending from the greater trochanter to the mid-thigh distally followed by a muscle

splitting incision. After nerve mobilisation, a transection injury was performed

(neurotmesis) immediately above the terminal nerve ramification using straight




microsurgical scissors. Rats were then randomly assigned to three experimental

groups. In one group (End-to-End), immediate cooptation with 7/0 monofilament nylon

epineurial sutures of the 2 transected nerve endings was performed, in a second group

(End-to-EndMemb) nerve transection was reconstructed by end-to-end suture, like in

the first group, and then enveloped by a membrane of equine collagen type III. In a

third group (End-to-EndMembCell) animals received the same treatment as the

previous group but with equine collagen type III membranes covered with neural cells

differentiated in vitro. Sciatic nerves from the contralateral site were left intact in all

groups and served as controls. To prevent autotomy, a deterrent substance was

applied to rats’ right foot (34; 35). The animals were intensively examined for signs of

autotomy and contracture during the postoperative and none presented severe

wounds, infections or contractures. All procedures were performed with the approval of

the Veterinary Authorities of Portugal in accordance with the European Communities

Council Directive of November 1986 (86/609/EEC).

6.3 Functional assessment

Evaluation of motor performance (EPT) and nociceptive function (WRL)

Motor performance and nociceptive function were evaluated by measuring extensor

postural thrust (EPT) and withdrawal reflex latency (WRL), respectively. The animals

were tested pre-operatively (week-0), at weeks 1, 2, and every two weeks thereafter

until week-20. The animals were gently handled, and tested in a quiet environment to

minimize stress levels. The EPT was originally proposed by Thalhammer and

collaborators, in 1995 (36) as a part of the neurological recovery evaluation in the rat

after sciatic nerve injury. For this test, the entire body of the rat, excepting the hind-

limbs, was wrapped in a surgical towel. Supporting the animal by the thorax and

lowering the affected hind-limb towards the platform of a digital balance, elicits the

EPT. As the animal is lowered to the platform, it extends the hind-limb, anticipating the

contact made by the distal metatarsus and digits. The force in grams (g) applied to the

digital platform balance (model TM 560; Gibertini, Milan, Italy) was recorded. The same

procedure was applied to the contralateral, unaffected limb. Each EPT test was

repeated 3 times and the average result was considered. The normal (unaffected limb)

EPT (NEPT) and experimental EPT (EEPT) values were incorporated into an equation

(Equation 8) to derive the functional deficit (varying between 0 and 1), as described by

Koka and Hadlock, in 2001 (37).



Motor Deficit = (NEPT – EEPT) / NEPT

Equation 8 EPT formula

To assess the nociceptive withdrawal reflex (WRL), the hotplate test was modified as

described by Masters and collaborators (38). The rat was wrapped in a surgical towel

above its waist and then positioned to stand with the affected hind paw on a hot plate

at 56ºC (model 35-D, IITC Life Science Instruments, Woodland Hill, CA). WRL is

defined as the time elapsed from the onset of hotplate contact to withdrawal of the hind

paw and measured with a stopwatch. Normal rats withdraw their paws from the

hotplate within 4.3 s or less (39). The affected limbs were tested 3 times, with an

interval of 2 min between consecutive tests to prevent sensitization, and the three

latencies were averaged to obtain a final result (40; 41). If there was no paw withdrawal

after 12 s, the heat stimulus was removed to prevent tissue damage, and the animal

was assigned the maximal WRL of 12 s (42).

Kinematic Analysis

Ankle kinematics during the stance phase of the rat walk was recorded prior nerve

injury (week-0), at week-2 and every 4 weeks during the 20-week follow-up time.

Animals walked on a Perspex track with length, width and height of respectively 120,

12, and 15 cm. In order to ensure locomotion in a straight direction, the width of the

apparatus was adjusted to the size of the rats during the experiments, and a darkened

cage was placed at the end of the corridor to attract the animals. The rats gait was

video recorded at a rate of 100 images per second (JVC GR-DVL9800, New Jersey,

USA). The camera was positioned perpendicular to the mid-point of the corridor length

at a 1-m distance thus obtaining a visualization field of 14-cm wide. Only walking trials

with stance phases lasting between 150 and 400 ms were considered for analysis,

since this corresponds to the normal walking velocity of the rat (20–60 cm/s) (42-44).

The video images were stored in a computer hard disk for latter analysis using an


Diego, USA). 2-D biomechanical analysis (sagittal plan) was carried out applying a two-

segment model of the ankle joint, adopted from the model firstly developed by Varejão

and collaborators (8; 42-44). Skin landmarks were tattooed at points in the proximal

edge of the tibia, in the lateral malleolus and, in the fifth metatarsal head. The rats’

ankle angle was determined using the scalar product between a vector representing

the foot and a vector representing the lower leg. With this model, positive and negative

values of position of the ankle joint indicate dorsiflexion and plantarflexion, respectively.

For each stance phase the following time points were identified: initial contact (IC),




opposite toe-off (OT), heel-rise (HR) and toe-off (TO) (8; 42-44), and were time

normalized for 100% of the stance phase. The normalized temporal parameters were

averaged over all recorded trials. Angular velocity of the ankle joint was also

determined where negative values correspond to dorsiflexion. Four steps were

analysed for each animal (8).

Histological and Stereological analysis

A 10-mm-long segment of the sciatic nerve distal to the site of lesion was removed,

fixed, and prepared for quantitative morphometry of myelinated nerve fibers. A 10-mm

segment of uninjured sciatic nerve was also withdrawn from control animals (N=6). The

harvested nerve segments were immersed immediately in a fixation solution containing

2.5% purified glutaraldehyde and 0.5% saccarose in 0.1M Sorensen phosphate buffer

for 6-8 hours. Specimens were processed for resin embedding as described in details

elsewhere (45; 46). Series of 2-µm thick semi-thin transverse sections were cut using a

Leica Ultracut UCT ultramicrotome (Leica Microsystems, Wetzlar, Germany) and

stained by Toluidine blue for stereological analysis of regenerated nerve fibers. The

slides were observed with a DM4000B microscope equipped with a DFC320 digital

camera and an IM50 image manager system (Leica Microsystems, Wetzlar, Germany).

One semi-thin section from each nerve was randomly selected and used for the

morpho-quantitative analysis. The total cross-sectional area of the nerve was

measured and sampling fields were then randomly selected using a protocol previously

described (46-48). Bias arising from the "edge effect" was coped with the employment

of a two-dimensional disector procedure which is based on sampling the "tops" of fibers

(49; 50). Mean fiber density in each disector was then calculated by dividing the

number of nerve fibers counted by the disector’s area (N/mm2). Finally, total fiber

number (N) in the nerve was estimated by multiplying the mean fiber density by the

total cross-sectional area of the whole nerve. Two-dimensional disector probes were

also used for the unbiased selection of a representative sample of myelinated nerve

fibers for estimating circle-fitting diameter and myelin thickness. Precision and

accuracy of the estimates were evaluated by calculating the coefficient of variation

(CV=SD/mean) and coefficient of error (CE=SEM/mean) (46-48).



7 Statistical analysis

Two-way mixed factorial ANOVA was used to test for the effect of time in the End-to-

End group (within subjects effect; 12 time points) and experimental groups (between

subjects effect, 3 groups). The sphericity assumption was evaluated by the Mauchly’s

test and when this test could not be computed or when sphericity assumption was

violated, adjustment of the degrees of freedom was done with the Greenhouse-

Geiser’s epsilon. When time main effect was significant (within subjects factor), simple

planned contrasts (General Linear Model, simple contrasts) were used to compare

pooled data across the three experimental groups along the recovery with data at

week-0. When a significant main effect of treatment existed (between subjects factor),

pairwise comparisions were carried out using the Tukey’s HSD test. At week-0,

kinematic data was recorded only from the End-to-End group so the main effect of time

was evaluated only in this group. Evaluation of the main effect of treatment on ankle

motion variables used only data after nerve injury. In this case, and when appropriate,

pairwise comparisons were made using the Tukey’s HSD test. Statistical comparisons

of stereological morpho-quantitative data on nerve fibers were accomplished with one-

way ANOVA test. Statistical significance was established as p<0.05. All statistical

procedures were performed by using the statistical package SPSS (version 14.0,

SPSS, Inc) except stereological data that were analysed using the software “Statistica

per discipline bio-mediche” (McGraw-Hill, Milan, Italy). All data in this study is

presented as mean ± standard error of the mean (SEM).




8 Results

8.1 Motor deficit and Nociception function

Motor deficit (EPT)

Before sciatic injury, EPT was similar in both hindlimbs in all experimental groups

(Figure 25). In the first week after sciatic nerve transection, near total EPT loss was

observed in the operated hindlimb, leading to a motor deficit ranging between 83 to

90%. The EPT response steadily improved during recovery but at week-20 the EPT

values of the injured side were still significantly lower compared to values at week-0

(p<0.05). A significant main effect for treatment was found [F(2,17) = 14.202; p=0.000],

with pairwise comparisons showing significantly better recovery of the EPT response in

the End-to-EndMembCell group when compared to the other two experimental groups

(p<0.05). At week-20, motor deficit decreased to 27% in the End-to-EndMembCell and

to 34% and 42% in the End-to-End and End-to-EndMemb groups, respectively (Figure

25).

EPT

0

20

40

60

80

100

0 1 2 4 6 8 10 12 14 16 18 20

W eek

Motor Deficit (%)

End-to-End End-to-EndMem b EndtoENdMem bCell

* #

Figure 25 - Weekly values of the percentage of motor deficit obtained by the Extensor

Postural Thrust (EPT) test. * Significantly different from week-0 all groups pooled together (p

< 0.05). # Group End-to-EndMembCell significantly different from the other groups (p < 0.05).

Results are presented as mean and standard error of the mean (SEM).



Nociception function (WRL)

In the week after sciatic transection, all the animals presented a severe loss of sensory

and nociception function acutely after sciatic nerve transection and the WRL test has to

be interrupted at the 12 s-cutoff time (Figure 26). During the following weeks there was

recovery in paw nociception which was more clearly seen between weeks 6 and 8

post-surgery. At week-6, half of the animals still had no withdrawal response to the

noxious thermal stimulus in the operated side, which is in contrast with week-8, when

all animals presented a consistent, although delayed, response. Despite such

improvement in WRL response, contrast analysis showed persistence of sensory deficit

in all groups by the end of the 20-weeks recovery time (p<0.05). No differences

between the groups was observed in the level of WRL impairment after the sciatic

nerve transection [F(2,17) = 1.563; p=0.238].

0

2

4

6

8

10

12

0 1 2 4 6 8 10 12 14 16 18 20

End-to-End End-to_EndMemb End-to_EndMembCell

* * * *

Figure 26 - Weekly values of the withdrawal reflex latency test. At week-1 all animals failed in responding to

the noxious thermal stimulus within the 12 sec cut-off time. No differences between the percentages of

motor deficit obtained by the Extensor Postural Thrust (EPT) test. * Significantly different from week-0 all

groups pooled together (p < 0.05). Results are presented as mean and standard error of the mean (SEM).




Kinematics Analysis

Figure 27 and Figure 28 display the mean plots, respectively for ankle joint angle and

ankle joint velocity during the stance phase of the rat walk. Comparisons to the normal

ankle motion can only be draw for the End-to-End group for reasons explained in the

Methods section. In the weeks following sciatic nerve transection, ankle joint motion

became severely abnormal, particularly throughout the second half of stance

corresponding to the push-off sub-phase. In clear contrast to the normal pattern of

ankle movement, at week-2 post-injury animals were unable to extend this joint and

dorsiflexion continued increasing during the entire stance, which is explained by the

paralysis of plantarflexor muscles. The pattern of the ankle joint motion seemed to

have improved only slightly during recovery. Contrast analysis was performed for each

of the kinematic parameters (Table 1 and Table 2) with somewhat different results. For

OT velocity and HR angle no differences existed before and after sciatic nerve

transection, whereas for OT angle differences from pre-injury values were significant

only at weeks 2 and 16 of recovery (p<0.05). The angle at IC showed a unique pattern

of changes, being unaffected at week-2 post-injury and altered from normal in the

following weeks of recovery. Probably the most consistent results are those of HR

velocity, TO angle and TO velocity. These parameters were affected immediately after

the nerve injury and remained abnormal along the entire 20-weeks recovery period

(p<0.05). The effect of the different tissue engineering strategies was assessed

comparing the kinematic data of the experimental groups only during the recovery

period (see Methods). Statistical analysis demonstrated that the collagen membrane

and the cells had no or little effect on ankle motion pattern recovery. Generally, no

differences in the kinematic parameters were found between the groups. Exceptions

were IC velocity in the End-to-EndMembCell group, which was different from the other

two groups (p<0.05), and OT angle in the End-to-EndMemb group that was also

different from the other two groups (p<0.05).



Figure 27- Kinematics plots in the sagittal plane for the angular position (°) of the ankle as it moves

through the stance phase, during the healing period of 20 weeks. The mean of each group is plotted




Figure 28 - Kinematics plots in the sagittal plane for the angular velocity (°/s) of the ankle as it moves

through the stance phase, during the healing period of 20 weeks. The mean of each group is plotted.



Table 1- Ankle kinematics and stance duration analysis were carried out prior to nerve injury (week-

0), at week-2, and every 4 weeks during the 20-week follow-up period.

Values of the ankle angular position (°) at initial contact (IC); opposite toe-off (OT); heel-rise (HR);

toe-off (TO) of the stance phase. Results are presented as mean and standard error of the mean

(SEM). N corresponds to the number of rats within the experimental group.




Table 2- Ankle kinematics and stance duration analysis were carried out prior to nerve injury (wek-

0), at week-2, and every 4 weeks during the 20-week follow-up period.

Values of the ankle angular velocity (°/sec) at initial contact (IC); opposite toe-off (OT); heel-rise

(HR); toe-off (TO) of the stance phase. Results are presented as mean and standard error of the

mean (SEM). N corresponds to the number of rats within the experimental group.



8.2 Histological and Stereological Analysis

Table 3- Stereological quantitative assessment density, total

number, diameter and myelin thickness of regenerated sciatic

nerve fibers at week-20 after neurotmesis. Values are presented as

mean ± SEM.




Figure 29 shows representative light micrographs of the regenerated sciatic nerves of

the three groups (Figure 29A-C) and control sciatic normal nerves (Figure 29D). As

expected, regeneration of axons was organized in many smaller fascicles in

comparison to controls. The results of the stereological analysis of myelinated nerve

fibers are reported in Table 3. Statistical analysis by ANOVA test revealed no

significant (p>0.05) difference regarding any of the morphological parameters

investigated in the regenerated axons from the three experimental groups. On the other

hand, comparison between regenerated and control nerves showed, as expected, the

presence of a significantly (p<0.05) higher density and total number of myelinated

axons in experimental groups accompanied by a significantly (p<0.05) lower fiber

diameter.

Figure 29 - Representative high resolution photomicrographs of nerve fibers form regenerated (A-C)

and normal (D) rat sciatic nerves. A: End-to-End. B:End-to-EndMemb. C:End-to-EndMembCell.

Magnification = × 1,500.



9 DISCUSSION

Transected peripheral nerves can regenerate provided that a connection is available

between the proximal and distal severed stumps and, when no substance loss

occurred; surgical treatment consists in direct end-to-end suturing of the nerve ends (2;

3). However, in spite of the progress of microsurgical nerve repair, the outcome of

nerve reconstruction is still far from being optimal (51). Since during regeneration

axons require neurotrophic support, they could benefit from the presence of a growth

factors delivery cell system capable of responding to stimuli of the local environment

during axonal regeneration.

In the present study, we aimed at investigating the effects of enwrapping the site of

end-to-end rat sciatic nerve repair with equine type III collagen nerve membranes

either alone or enriched with N1E-115 pre-differentiated into neural cells in the

presence of 1.5% of DMSO. The rationale for the utilization of the N1E-115 cells was to

take advantages of the properties of these cells as a neural-like cellular source of

neurotrophic factors (6-10).

Results showed that enwrapment with a collagen membrane, with or without neural cell

enrichment, did not lead to any significant improvement in most of functional and

stereological predictors of nerve regeneration that we have assessed. The only

exception was represented by motor deficit recovery which was significantly improved

after lesion site enwrapment with membrane enriched with neural cells pre-

differentiated from N1E-115 cell line.

Natural tissues possess several advantages when compared to synthetic materials,

when use to reconstruct peripheral nerves after injury. Natural materials are more likely

to be biocompatible than artificial materials, are less toxic, and provide a support

structure to promote cell adhesion and migration. Drawbacks, on the other hand,

include potential difficulties with isolation and controlled scale-up. In addition to intact

acellular tissues, a great deal of research has focused on the use of purified natural

ECM proteins and glycosaminoglycans, which can be modified to serve as appropriate

scaffolding. ECM molecules, such as laminin, collagen, and fibronectin, have been

shown to play a significant role in axonal development and repair in the body (24; 52).

There are a number of examples in which the ECM proteins laminin, fibronectin, and

collagen have been used for nerve repair applications (21; 23-25; 52-56). For example,

silicone tubes filled with laminin, fibronectin, and collagen show improved regeneration

over a 10 mm rat sciatic nerve gap compared to empty silicone controls (14). Collagen




filaments have also been used to guide regenerating axons across 20–30 mm defects

in rats (25; 26; 57). Further studies have shown that oriented fibers of collagen within

gels, aligned using magnetic fields, provide an improved template for neurite extension

compared to randomly oriented collagen fibers (22; 27). Rates of regeneration after

neurotmesis comparable to those using a nerve autograft have been achieved using

collagen tubes containing a porous collagen-glycosaminoglycan matrix (29; 30).

Results of this study contribute to the lively debate about the employment of cell

transplantation for improving post-traumatic nerve regeneration (58; 59). Actually, a

great enthusiasm among researchers and especially the public opinion has risen over

the last years about cell-based therapies in Regenerative Medicine (60-62) and there

seems to be widespread conviction that this type of therapy is not only effective but

also very safe in comparison to other pharmacological or surgical therapeutic

approaches. By contrast, recent studies showed that cell-based therapy might be

ineffective for improving nerve regeneration [66-69], and results of the present study

are in line with these observations. Recently, it has even been shown that N1E-115 cell

transplantation can also have negative results by hindering the nerve regeneration

process after tubulisation repair (17). Of course, the choice of the cell type to be used

for transplantation is very important for the therapeutic success and use of another cell

type could have led to better results, especially when the cellular system of choice is

derived from autologous or heterologous stem cells1 (17-20; 58; 63). Moreover, the

construction of more appropriate tube-guides with integrated growth factor delivery

systems and/or cellular components could improve the effectiveness of nerve tissue

engineering. In fact, single-molded tube guides may not give sufficient control over both

the mechanical properties and the delivery of bioactive agents. More complex devices

will be needed, such as multilayered tube guides where growth factors are entrapped in

polymer layers with varying physicochemical properties or tissue engineered tube

guides containing viable stem cells (17-20; 58; 63). The combination of two or more

growth factors will likely exert a synergistic effect on nerve regeneration, especially

when the growth factors belong to different families and act via different mechanisms.

Combinations of growth factors can be expected to enhance further nerve

regeneration, particularly when each of them is delivered at individually tailored kinetics

(16-20; 58; 63). The determination and control of suitable delivery kinetics for each of

several growth factors will constitute a major hurdle both technically and biologically

with the biological hurdle lying in the compliance with the naturally occurring cross talk

between growth factors and cells. A solution to this problem may be the use of



autologous stem cells because they can synthesize several growth factors and

differentiate into Schwann cells which are critical for very long gaps (16-20; 58; 63).

Previous work already published by other research groups, point out a very interesting

source of stem cells for nerve regeneration of peripheral nerve and spinal cord. They

developed hair follicle pluripotent stem cells (hfPS) and have shown that these cells

can differentiate to neurons, glial cells in vitro, and other cell types, and can promote

nerve and spinal cord regeneration in vivo. These cells are located above the hair

follicle bulge (hfPS cell area) and are nestin and CD34 positive, and keratin 15

negative (64-67). The mouse hfPS cells were implanted into the gap region of the

severed sciatic and tibial nerve of mice. These cells, after 6-8 weeks,

transdifferentiated largely into Schwann cells. Also, blood vessels formed a network

around the joined sciatic and tibial nerve. Function of the rejoined sciatic and tibial

nerve was confirmed by contraction of the gastrocnemius muscle upon electrical

stimulation and by walking track analysis (64-66). hfPS cells can promote axonal

growth and functional recovery after peripheral nerve injury, offering an important

opportunity for future clinical application. These hfPS cells, in contrast to Embrionic

stem cells, N1E-115 cells after in vitro differentiation and induced pluripotent stem

cells, do not require any genetic manipulation, are readily accessible from any patient,

and lack the ethical issues, do not form tumors.




References

1. Lundborg G. Alternatives to autologous nerve grafts. [Internet]. Handchirurgie,

Mikrochirurgie, plastische Chirurgie. 2004 Feb ;36(1):1-7.

2. Matsuyama T, Mackay M, Midha R. Peripheral nerve repair and grafting

techniques: a review [Internet]. Neurologia medico-chirurgica. 2000 ;40(4):187–

199.

3. Siemionow M, Brzezicki G. Chapter 8: Current techniques and concepts in

peripheral nerve repair. [Internet]. International review of neurobiology. 2009 Jan

;87141-72.



Academy of Sciences of the United States of America. 1976 Feb ;73(2):462-6.

5. Prasad KN, Kentroti S, Edwards-Prasad J, Vernadakis A, Imam M, Carvalho E,

et al. Modification of the expression of adenosine 3ʼ,5'-cyclic monophosphate-

induced differentiated functions in neuroblastoma cells by beta-carotene and D-

alpha-tocopheryl succinate. [Internet]. Journal of the American College of

Nutrition. 1994 Jun ;13(3):298-303.




;18(5):323-8.








methods. 2007 Jun ;163(1):92-104.




;28(6):458-70.



10. Amado S, Simões MJ, Armada da Silva PAS, Luís AL, Shirosaki Y, Lopes MA,

et al. Use of hybrid chitosan membranes and N1E-115 cells for promoting nerve

regeneration in an axonotmesis rat model. [Internet]. Biomaterials. 2008 Nov

;29(33):4409-19.

11. Doolabh VB, Hertl MC, Mackinnon SE. The role of conduits in nerve repair: a

review. [Internet]. Reviews in the neurosciences. 7(1):47-84.

12. Jensen JN, Tung THH, Mackinnon SE, Brenner MJ, Hunter DA. Use of anti-

CD40 ligand monoclonal antibody as antirejection therapy in a murine peripheral

nerve allograft model. [Internet]. Microsurgery. 2004 Jan ;24(4):309-15.



[Internet]. Brain research reviews. 2006 Sep ;52(2):381-8.

14. Chen YS, Hsieh CL, Tsai CC, Chen TH, Cheng WC, Hu CL, et al. Peripheral

nerve regeneration using silicone rubber chambers filled with collagen, laminin

and fibronectin. [Internet]. Biomaterials. 2000 Aug ;21(15):1541-7.

15. Dunnen WF den, Robinson PH, Wessel R van, Pennings AJ, Leeuwen MB van,

Schakenraad JM. Long-term evaluation of degradation and foreign-body reaction

of subcutaneously implanted poly(DL-lactide-epsilon-caprolactone). [Internet].

Journal of biomedical materials research. 1997 Sep ;36(3):337-46.



;5293-347.




Jun ;14(6):979-93.

18. Battiston B, Raimondo S, Tos P, Gaidano V, Audisio C, Scevola A, et al.

Chapter 11: Tissue engineering of peripheral nerves [Internet]. Elsevier; 2009.

19. Zacchigna S, Giacca M. Chapter 20: Gene therapy perspectives for nerve repair.

[Internet]. International review of neurobiology. 2009 Jan ;87381-92.

20. Karagoz H, Ulkur E, Uygur F, Senol MG, Yapar M, Turan P, et al. Comparison of

regeneration results of prefabricated nerve graft, autogenous nerve graft, and

vein graft in repair of nerve defects. [Internet]. Microsurgery. 2009 Jan

;29(2):138-43.




21. Kauppila T, Jyväsjärvi E, Huopaniemi T, Hujanen E, Liesi P. A laminin graft

replaces neurorrhaphy in the restorative surgery of the rat sciatic nerve.

[Internet]. Experimental neurology. 1993 Oct ;123(2):181-91.

22. Ceballos D, Navarro X, Dubey N, Wendelschafer-Crabb G, Kennedy WR,

Tranquillo RT. Magnetically aligned collagen gel filling a collagen nerve guide

improves peripheral nerve regeneration. [Internet]. Experimental neurology.

1999 Aug ;158(2):290-300.

23. Toba T, Nakamura T, Lynn AK, Matsumoto K, Fukuda S, Yoshitani M, et al.

Evaluation of peripheral nerve regeneration across an 80-mm gap using a

polyglycolic acid (PGA)--collagen nerve conduit filled with laminin-soaked

collagen sponge in dogs. [Internet]. The International journal of artificial organs.

2002 Mar ;25(3):230-7.

24. Grimpe B, Silver J. The extracellular matrix in axon regeneration. [Internet].

Progress in brain research. 2002 Jan ;137333-49.

25. Yoshii S, Oka M. Peripheral nerve regeneration along collagen filaments.

[Internet]. Brain research. 2001 Jan ;888(1):158-162.

26. Itoh S, Takakuda K, Kawabata S, Aso Y, Kasai K, Itoh H, et al. Evaluation of

cross-linking procedures of collagen tubes used in peripheral nerve repair.

[Internet]. Biomaterials. 2002 Dec ;23(23):4475-81.

27. Dubey N, Letourneau PC, Tranquillo RT. Guided neurite elongation and

schwann cell invasion into magnetically aligned collagen in simulated peripheral

nerve regeneration. [Internet]. Experimental neurology. 1999 Aug ;158(2):338-

50.

28. Archibald SJ, Shefner J, Krarup C, Madison RD. Monkey median nerve repaired

by nerve graft or collagen nerve guide tube. [Internet]. The Journal of

neuroscience : the official journal of the Society for Neuroscience. 1995 May

;15(5 Pt 2):4109-23.

29. Chamberlain LJ, Yannas IV, Hsu HP, Strichartz G, Spector M. Collagen-GAG

substrate enhances the quality of nerve regeneration through collagen tubes up

to level of autograft. [Internet]. Experimental neurology. 1998 Dec ;154(2):315-

29.

30. Chamberlain LJ, Yannas IV, Hsu HP, Strichartz GR, Spector M. Near-terminus

axonal structure and function following rat sciatic nerve regeneration through a

collagen-GAG matrix in a ten-millimeter gap. [Internet]. Journal of neuroscience

research. 2000 Jun ;60(5):666-77.



31. Fu Y. Contributing Factors to Poor Functional Nerve Repair : Prolonged

Denervation Recovery after Delayed. 1995 ;15(May):3886-3895.

32. Tria MA, Fusco M, Vantini G, Mariot R. Pharmacokinetics of nerve growth factor

(NGF) following different routes of administration to adult rats. [Internet].

Experimental neurology. 1994 Jun ;127(2):178-83.





of America. 1972 Jan ;69(1):258-63.



;47(1):31-9.








nerve transection. [Internet]. Experimental neurology. 2001 Mar ;168(1):192-5.



biodegradable polymer matrix. [Internet]. Anesthesiology. 1993 Aug ;79(2):340-

6.


after sciatic nerve blockade. [Internet]. Anesthesiology. 1997 Apr ;86(4):957-65.

40. Campbell JN. Nerve lesions and the generation of pain. [Internet]. Muscle &

nerve. 2001 Oct ;24(10):1261-73.

41. Shir Y, Campbell JN, Raja SN, Seltzer Z. The correlation between dietary soy

phytoestrogens and neuropathic pain behavior in rats after partial denervation.

[Internet]. Anesthesia and analgesia. 2002 Feb ;94(2):421-6.










2003 Jan ;23(4):346-53.



nerve. 2002 Nov ;26(5):630-5.




502.




review of neurobiology. 2009 Jan ;8781-103.



myelinated nerve fibers in peripheral nerves. [Internet]. Annals of anatomy. 2000

Jan ;182(1):23-34.

48. Geuna S, Tos P, Guglielmone R, Battiston B, Giacobini-Robecchi MG.

Methodological issues in size estimation of myelinated nerve fibers in peripheral

nerves. [Internet]. Anatomy and embryology. 2001 Jul ;204(1):1-10.



Jan ;24(1):72-6.

50. Geuna S. The revolution of counting "tops": two decades of the disector principle

in morphological research. [Internet]. Microscopy research and technique. 2005

Apr ;66(5):270-4.

51. Gordon T, Sulaiman O, Boyd JG. Experimental strategies to promote functional

recovery after peripheral nerve injuries. [Internet]. Journal of the peripheral

nervous system : JPNS. 2003 Dec ;8(4):236-50.

52. Rutishauser U. Adhesion molecules of the nervous system. [Internet]. Current

opinion in neurobiology. 1993 Oct ;3(5):709-15.






137.

54. Tong XJ, Hirai K, Shimada H, Mizutani Y, Izumi T, Toda N, et al. Sciatic nerve

regeneration navigated by laminin-fibronectin double coated biodegradable

collagen grafts in rats. [Internet]. Brain research. 1994 Nov ;663(1):155-62.

55. Whitworth IH, Brown RA, Doré C, Green CJ, Terenghi G. Orientated mats of

fibronectin as a conduit material for use in peripheral nerve repair. [Internet].

Journal of hand surgery (Edinburgh, Scotland). 1995 Aug ;20(4):429-36.

56. Ahmed Z, Brown RA. Adhesion, alignment, and migration of cultured Schwann

cells on ultrathin fibronectin fibres. [Internet]. Cell motility and the cytoskeleton.

1999 Jan ;42(4):331-43.

57. Yoshii S, Oka M, Shima M, Taniguchi A, Akagi M. 30 mm regeneration of rat

sciatic nerve along collagen filaments. [Internet]. Brain research. 2002 Sep

;949(1-2):202-8.

58. Terenghi G, Wiberg M, Kingham PJ. Chapter 21: Use of stem cells for improving

nerve regeneration. [Internet]. International review of neurobiology. 2009 Jan

;87393-403.

59. Lundborg G. Enhancing posttraumatic nerve regeneration. [Internet]. Journal of

the peripheral nervous system : JPNS. 2002 Oct ;7(3):139-40.[cited 2011 Mar

31] Available from: http://www.ncbi.nlm.nih.gov/pubmed/12365560

60. Tohill M, Terenghi G. Stem-cell plasticity and therapy for injuries of the

peripheral nervous system. [Internet]. Biotechnology and applied biochemistry.

2004 Aug ;40(Pt 1):17-24.

61. Pfister LA, Papaloïzos M, Merkle HP, Gander B. Nerve conduits and growth

factor delivery in peripheral nerve repair. [Internet]. Journal of the peripheral

nervous system : JPNS. 2007 Jul ;12(2):65-82.

62. Chalfoun CT, Wirth GA, Evans GRD. Tissue engineered nerve constructs: where

do we stand? [Internet]. Journal of cellular and molecular medicine. 10(2):309-

17.

63. Dahlin L, Johansson F, Lindwall C, Kanje M. Chapter 28: Future perspective in

peripheral nerve reconstruction. [Internet]. Elsevier; 2009. [cited 2010 Nov 22]

Available from: http://www.ncbi.nlm.nih.gov/pubmed/19682657

64. Amoh Y, Kanoh M, Niiyama S, Hamada Y, Kawahara K, Sato Y, et al. Human

hair follicle pluripotent stem (hfPS) cells promote regeneration of peripheral-

nerve injury: an advantageous alternative to ES and iPS cells. [Internet]. Journal

of cellular biochemistry. 2009 Aug ;107(5):1016-20.




65. Amoh Y, Li L, Campillo R, Kawahara K, Katsuoka K, Penman S, et al. Implanted

hair follicle stem cells form Schwann cells that support repair of severed

peripheral nerves. [Internet]. Proceedings of the National Academy of Sciences

of the United States of America. 2005 Dec ;102(49):17734-8.

66. Amoh Y, Li L, Katsuoka K, Penman S, Hoffman RM. Multipotent nestin-positive,

keratin-negative hair-follicle bulge stem cells can form neurons. [Internet].

Proceedings of the National Academy of Sciences of the United States of

America. 2005 May ;102(15):5530-4.

67. Li L, Mignone J, Yang M, Matic M, Penman S, Enikolopov G, et al. Nestin

expression in hair follicle sheath progenitor cells. [Internet]. Proceedings of the

National Academy of Sciences of the United States of America. 2003 Aug

;100(17):9958-61.

Chapter 5 - The sensitivity of two-dimensional hindlimb

joints kinematics analysis in assessing function in rats

after sciatic nerve crush

Amado S, Armada-da-Silva P, João F, Mauricio AC, Luis AL, Simões MJ, et al. The

sensitivity of two-dimensional hindlimb joints kinematics analysis in assessing function

in rats after sciatic nerve crush. Behavioural brain research. 2011; submitted.



1 Introduction

Injury of peripheral nerves is generally followed by abnormal nerve regeneration,

axonal misrouting, and incorrect reinnervation of the target organs (1; 2). Intense

investigation exists trying to develop new nerve treatments and neurorehabilitation

strategies, including alternative surgical techniques (3; 4), the use of biomaterial and

cellular systems (5-7), nerve electrical stimulation (8) and different modalities of

exercise (9; 10).

The evaluation of the effectiveness of peripheral nerve treating strategies, however,

requires accurate assessment of functional recovery, since the latter is the major goal

of peripheral nerve treatment and the most important outcome when translating

experimental treatment approaches to humans. Notwithstanding, functional

assessment in animal models of peripheral nerve injury is a challenging task. The rat

sciatic nerve is a widely used model of peripheral nerve injury. In this case, computer-

assisted biomechanical analysis of joint motion during walking using video recordings

and focusing on ankle joint motion as been shown to be a valid and reliable method to

evaluate functional recovery and the reestablishment of proper reinnervation (11-13).

When compared to other functional tests, in particular the standard sciatic functional

index test, analysis of ankle motion pattern during walking in sciatic-injured rats

demonstrates higher sensitivity to motion deficits and shows better relationship with the

degree of nerve regeneration (13). Also, recent studies demonstrate that measures of

two-dimensional (2D) ankle joint motion during rat walking present good day-to-day

and inter-observer reliabilities and are capable of identifying animals carrying injuries to

the sciatic, tibial or common peroneal nerves (14).

Joint kinematics during a given movement, such as walking, is usually represented by

a trajectory in an angle versus time space. Specific kinematic parameters are then

extracted from such trajectories considering particular events that can be

unambiguously defined (12; 14; 15). This can result in a varied number of variables that

may be used to measure changes in the normal pattern of joint motion. In the case of

experimental injury to the sciatic nerve of the rat, authors have focused on the ankle

joint kinematics during walking, based on the fact this nerve innervates the muscles

crossing this joint (11; 12; 14-16). Some studies limited their walking analyses only to

the stance phase, based on the argument that muscles are particularly active during

the stance phase in order to support the load of body weight and to generate the power

needed to accelerate the body forwardly (5; 11; 12; 17). However, more recent studies

have choose to analyze ankle motion during both the stance and swing phases of

Chapter 5 - The sensitivity of two-dimensional hindlimb joints kinematics analysis in assessing function in rats after

sciatic nerve crush 107


walking (14; 18), thus raising the ability to unveil motion deficits of the ankle joint after

sciatic nerve injury since this nerve supplies both dorsiflexor and plantarflexor muscles.

Also, 2D video analysis of ankle joint motion in the two walking phases is important to

distinguish between injuries to the sciatic, tibial or common peroneal nerves (18).

In previous studies we also conducted video recordings of the rat walk in the sagittal

plane pre and post sciatic crush injury, measured ankle joint angles at given time points

during the stance phase and calculated the angular velocities at these instants of time

(5; 11). The angular velocity of the ankle joint motion was used as an attempt to

improve the accuracy of ankle kinematic analysis to detect walking dysfunction. In fact,

there are cases where results of ankle joint angles are difficult to interpret, raising

concerns about their validity as a functional assessment. For example in our previous

study (11), we reported in sciatic nerve-crushed rats unchanged ankle angle at the

instant of opposite toe off in the early paralytic stage (two weeks post-injury) and

altered ankle angle at this time point thereafter and until the end of the twelve weeks

follow-up time. Inconsistent results were found also for other ankle kinematics

parameters in this study (11), as well as in other studies (14; 19). These suggest that

not all ankle kinematic parameters are sensitive to severe functional deficit caused by

complete sciatic nerve denervation and therefore are not a valid functional test.

Importantly, the use of inaccurate kinematic parameters blurs the results and may

become an obstacle to draw sound conclusions.

Sciatic nerve crush is also likely to affect the kinematics of hindlimb joints other than

the ankle joint. Motion changes at the hip and knee joints might accompany ankle joint

dysfunction during walking either to compensate for the abnormal movement of the

ankle joint or else due to higher-level reorganization of the whole hindlimb action in

response to long-lasting motor and proprioceptive deficits of the muscles supplied by

the sciatic nerve (20). Therefore, it is possible that screening hip and knee motion will

enhance the sensitivity and specificity of kinematic analysis of walking in the rat sciatic

nerve model. To our knowledge a walking analysis that includes a description of the

hip, knee and ankle joint kinematics after sciatic nerve injury has never been

performed.

The sensitivity (i.e. the true positive rate) and specificity (i.e. the true negative rate) are

two key properties of diagnostic and screening tests. These test properties can be

evaluated using linear discriminant analysis particularly if we are talking of multivariate

tests that combine several parameters (21). Discriminant analysis simply addresses

classification or discrimination in which objects or patterns are classified into one of

several distinct populations using a predictive model developed on a set of



independent variables (22). The number of subjects that are correctly classified by the

discriminant model provides a measure of the sensitivity of the set of testing variables

used in its building. Conversely, the number of subjects misclassified is a measure of

the specificity of the test variables. Additionally, discriminant analysis gives information

about the relative importance of each of the independent variables in their classification

role, yielding a sound basis for discarding variables with little contribution in separating

the groups. Therefore, discriminant analysis may be used to identify which are the best

walking joint kinematic parameters to assess functional recovery after sciatic nerve

injury in the rat.

In this study we conduct a detailed 2D biomechanical analysis of the hip, knee and

ankle joints in the rat and employed linear discriminant analysis to assess the

sensitivity and specificity of a large set of spatiotemporal and joint kinematic variables

as a test of movement dysfunction after sciatic nerve injury. Sciatic nerve-crushed rats

in the denervated (one week post injury) and in the reinnervated (twelve weeks post

injury) phases were used as they may serve as a model of animals with severe and

minor walking deficits, respectively. Sham-operated controls were used as the normal

walking group.




2 Methods and Materials

Twenty four male Sprague-Dawley rats (Harlan Laboratories, Udine, Italy) were

randomly assigned to one of three groups. A crush injury to the right sciatic nerve was

induced in animals in two of the groups, while animals of the third group underwent a

sham surgery (Sham). Walking tests were performed one week after sciatic nerve

injury, corresponding to a period of complete denervation and paralysis of sciatic-

innervated muscles (DEN group) (23), and twelve weeks after either sciatic nerve crush

(REINN group) or sham surgery (Sham group). After twelve weeks from a sciatic nerve

crush, muscle reinnervation has occurred and functional recovery based on the

standard sciatic functional index is complete (11; 24; 25). As a result of the longer

follow up time, mean (±SD) weight was higher in the REINN (454.1±15.8 g) and Sham

(435.5±17.6) groups, compared to the DEN group (351±5.1 g). All procedures were

performed with the approval of the Veterinary Authorities of Portugal in accordance

with the European Communities Council Directive of November 1986 (86/609/EEC).

The sciatic crush injury procedure was described in detail elsewhere (11). Briefly,

under deep anaesthesia (ketamine 9 mg/100 g; xylazine 1.25 mg/100 g, i.p.), the right

sciatic nerve was exposed by skin incision and by splitting the fascia and muscles

overlying the nerve. The right sciatic nerve was crushed in the gluteal region, well

before its trifurcation point, by applying pressure for 30 s using a non-serrated clamp

(25; 26). The muscle, fascia, and skin were closed with 4/0 resorbable sutures. To

prevent autotomy, a deterrent substance was applied to the rats’ right hindleg and foot

[(27; 28). The animals were intensively examined for signs of autotomy and contracture

and none of the animals presented severe wounds (absence of a part of the foot or

severe infection) or contractures during the study. In sham-operated animals, skin and

muscle incisions were performed and the sciatic nerve exposed and mobilized, but no

injury was induced to this nerve. All animals were left to recover in their cages with no

other measures taken that could enhance nerve regeneration and/or functional

recovery.

2.1 Kinematic analysis

An optoelectronic system of six infrared cameras (Oqus-300, Qualisys, Sweden)

operating at a frame rate of 200Hz was used to record the motion of the right hindlimb.

After shaving, seven reflective markers with 2mm diameter were attached to the skin of

the right hindlimb at the following bony prominences: (1) tip of fourth finger, (2) head of



fifth metatarsal, (3) lateral malleolus, (4) lateral knee joint, (5) trochanter major, (6)

anterior superior iliac spine, and (7) ischial tuberosity. All markers were placed by the

same person. Animals walked on a Perspex track with length, width and height of

respectively 120, 12 and 15cm with two darkened cages placed at both ends of the

corridor to attract the animals and facilitate walking. Before data collection, all animals

performed two or three conditioning trials to be familiarized with the corridor. Cameras

were positioned to minimize light reflection artifacts and to allow recording 4 to 5

consecutive walking cycles, defined as the time between two consecutive initial ground

contacts (IC) of the right fourth finger. The motion capture space was calibrated

regularly using a fixed set of markers and a wand of known length (20 cm) moved

across the recorded field. Calibration was accepted when the standard deviation of the

wand’s length measures was below 0.4 mm. Planar motion in the sagittal plane of the

hip, knee and ankle joint was calculated with Visual 3D software (C-Motion, Inc,

Germantown, USA) by a computational procedure implementing the dot product

between the skeletal segments articulated by these joints. Joint velocity was also

calculated using the first derivative of joint angle motion. The trajectory of the reflective

markers was smoothed using a Butterworth low-pass filter with a 6 Hz cut-off. Data

were obtained by averaging six walking cycles. Each walking cycle was time

normalized by interpolation using the longest trial. Our angles convention was: 1)

decreasing hip angle value – hip flexion; 2) decreasing knee angle value – knee

flexion; 3) positive ankle angle value – dorsiflexion. Positive angular velocity indicates

increasing angle values. This also applies to the ankle joint: positive angular velocity

signifies increasing dorsiflexion. Joint angle and angular velocity values were

measured at the instants of initial ground contact (IC), defined as the time point where

horizontal velocity of the toe reflective marker becomes zero, midstance (MSt), the

point of fifty per cent duration of the stance phase, toe-off (TO), defined as the instant

the horizontal velocity of the toe’s reflective marker increases from zero, and midswing

(MSw), defined as the point of fifty per cent duration of the swing phase. Additional

kinematics parameters recorded included peak angles and angular velocities during

both phases of the walking cycle.

Spatiotemporal parameters, including walking velocity, walking cycle duration, stride

length, stance duration, and swing duration, were directly given by Visual 3D software

based on horizontal displacement of the reflective markers.

Interjoint coordination shape patterns were studied by using cyclograms or else

designated angle-angle plots. The angle-angle plots included those of hip-knee, knee-

ankle and hip-ankle. Cyclograms perimeters were measured using the euclidean

distance between each two consecutive points.




3 Statistical analysis and modelling

Univariate ANOVA was used to test for differences between the groups. Further

pairwise comparisons were performed using the Tukey’s HSD test.

Linear disciriminant analysis was used to predict walking function and to classify

animals after sciatic nerve crush and controls. Four models were tested utilizing

spatiotemporal data and joint kinematics data of the hip, knee and ankle joints,

separately. For each subset all parameters were employed, meaning using both joint

angle and angular velocity in the cases of joint kinematics. Peak angles and angular

velocities were excluded from this analysis. The stepwise method was selected to input

the criterion variables with the Wilk’s Lambda choose as the method for inclusion or

rejection of the predictive variables. The leave-one-method was implemented for

further model cross-validation. This method is particularly useful when using small

samples and when a training sub-sample is not available for unbiased model

development. This method operates by having each subject classified by a discriminant

model constructed on the basis of data pertaining to the remaining subjects. Data are

presented as mean and standard deviation (SD). Statistical analysis was performed

using SPSS software package (version 17, SPSS Inc). Statistical significance was

accepted at p<0.05.



4 Results

4.1 Spatiotemporal parameters

Spatiotemporal data are presented in Table 4. The DEN group had a lower walking

velocity and a shorter stride length compared to the Sham group (p=0.005 and

p=0.007, for walking velocity and stride length, respectively) but not to the REINN

group (p=0.058 and p=0.145, for walking velocity and stride length, respectively). The

walking cycle time was longer in the DEN group, again only compared to the Sham

group (p=0.046, Sham group; p=0.131, REINN group). The DEN group also showed a

significantly longer swing duration, compared to both REINN (p=0.000) and Sham

(p=0.000) groups. No differences in spatiotemporal measures were observed between

the REINN and the Sham groups.

Table 4 - Mean values for spatiotemporal data for each experimental group; eight animals per

group

Groups Walking velocity

(ms-1)

Cycle duration

(s)

Stride length (m) Stance time (s) Swing time (s)

DEN 0.161±0.029b 0.852±0.057b 0.133±0.003b 0.238±0.030 0.175±0.009a

REINN 0.210±0.016 0.730±0.035 0.148±0.006 0.258±0.010 0.123±0.002

Sham 0.231±0.014 0.697±0.030 0.160±0.005 0.248±0.015 0.125±0.002

aSignificantly different from REINN and Sham; p<0.05;

bSignificantly different from Sham only; p<0.05




4.2 Joint kinematics

Plots of mean ankle, hip and knee joint angle and angular velocity are shown in Figure

30 with the parameter measures reported in Table 5, Table 6 and Table 7

The ankle was the most affected joint after sciatic nerve injury. During the stance

phase, the most evident ankle motion change is the higher overall dorsiflexion angle in

Figure 30 - Mean trajectories for hip (upper graphs), knee (middle graphs) and ankle (lower

graphs) joint angle (left hand graphs) and angular velocity (right hand graphs) during the rat’s

walking cycle. Full lines correspond to the stance phase and dashed lines correspond to the

swing phase of the walking cycle. In the hip joint angle, increased extension during the stance

phase is visible in the DEN group (single line) as well as the increased flexion in the REINN

group (triple line). Also evident, the deep altered ankle kinematics in both stance and swing

phases in the DEN group and similar trajectories of ankle angle and angular velocity between

the REINN and Sham groups. All groups n=8.



the DEN group. In this group, ankle motion is also severely affected across the swing

phase, and the characteristic movement of fast dorsiflexion followed by plantarflexion

that can be seen in REINN and Sham groups is absent in the DEN group. Accordingly,

changes in ankle angular velocity in the DEN group occur mostly during the swing

phase and not during the stance phase. In both REINN and Sham groups, ankle’s

maximal dorsiflexion and plantarflexion angular velocities are recorded during the

swing phase, that obviously correspond to a large range of movement of the ankle

performed in a short time interval. In the DEN group, these two peak angular velocities

are no longer observed and in replace of a large amplitude movement, the ankle

performs a slow and small amplitude plantarflexion movement along the entire swing

phase. This ankle motion pattern is consistent with total paralysis of the hindleg

muscles and loss of contractile force of both the dorsiflexor and plantarflexor muscles

caused by complete injury of the motoneuron axons in the sciatic nerve.

Reflecting the deep ankle kinematics changes in the DEN group, ANOVA reveals

significant differences between the groups for all ankle kinematics parameters,

excepting ankle’s angular velocity at MSw (p=0.192, Table 5). Pairwise tests showed

significant differences in ankle angles and angular velocities between the DEN group

and both REINN and Sham groups, excepting for ankle angle at MSw, where the

differences were significant only between DEN and Sham groups (11.4±11.3º and

28.1±13.9º, for DEN and Sham groups, respectively, p<0.05). Peak dorsiflexion angle

during the stance phase was increased in the DEN group comparing with the other two

groups (47.42±7.82º, 3.60±14.19º, 4.53±15.19º, DEN, REINN, Sham groups,

respectively; both pairwise comparisons, p=0.000). No clear peak plantarflexion angle

could be identified in animals of the DEN group, either in the stance or in the swing

phase. Ankle peak angles and angular velocities during the swing phase in REINN and

Sham groups were similar (Table 5).

Hip joint motion was also affected by sciatic nerve crush (Figure 30 and Table 6). Hip

joint excursion during walking was considerably increased in the DEN group. This

increased range of motion was due to a larger hip extension during the stance phase

while hip flexion during the swing phase remained relatively unaffected (Figure 30).

Accordingly, hip angular velocity was also higher in the DEN group, than in the REINN

and Sham groups (Figure 30). In the DEN group, hip angle at IC, TO, and MSt were

significantly different compared to the Sham group (p<0.05). Hip joint angle and

angular velocity at MSt were different between the REINN and Sham groups (p=0.046

and p=0.034 for hip angle and velocity, respectively). Hip’s peak extension angle

occurred at around TO and therefore it was not computed. In the swing phase, a

distinct hip flexion peak angle could be noticed only in the DEN group. However, its




value was similar to maximum hip flexion angle in the other groups for the same

walking phase (Figure 30). Hip peak flexion and extension velocities were increased in

DEN group compared to both REINN and Sham groups.

The changes in knee joint motion with denervation (DEN group) were less pronounced

than those registered in ankle and hip joints (Figure 30). The most evident knee angle

change in the denervated group is increased knee angle extension, particularly during

the stance phase (Table 7). The increased knee extension in the DEN group is likely a

compensation for the augmented ankle dorsiflexion and a strategy to maintain the

whole limb length. Knee angle changes were also noticed in the REINN group. This

group showed increased knee flexion at TO (50.29±15.09º), compared to both Sham

(72.50±11.26º, p=0.037) and DEN (79.40±13.55º, p=0.001) groups. Also, peak knee

flexion angle during the swing phase was increased in REINN group, compared to the

DEN group (p=0.047). This reflected a global increase in knee flexion throughout the

entire walking cycle in the REINN group compared to knee motion in sham-control

animals (Figure 30). No differences in knee angle at IC and MSw were found between

the groups. Knee angular velocity in the DEN group was different from either the Sham

or REINN groups at MSt, TO, and MSw. No differences in knee angular velocity were

found between REINN and Sham groups. Knee peak flexion velocity during the swing

phase was similar in all groups.

Table 5 - Mean values for ankle joint angle and angular velocity for each experimental group; eight animals per group

Ankle joint

Groups IC MSt

Peak value

extension

(plantarflexion)

Peak value flexion

(dorsiflexion) TO

Peak value

extension

(plantarflexion)

Peak value

flexion

(dorsiflexion)

MSw

Ang

le (

º)

DEN -0.6±

12.9a

46.2±

8.4a

-0.81±

13.26

47.42±

7.82a

17.1±

9.4a

-5.31±

12.65b

17.50±

9.57

11.4±

11.3b

REINN -20.1±

12.5

4.2±

14.1

-22.45±

13.08

3.60±

14.19

-14.6±

13.6

-20.86±

12.69

22.63±

14.89

23.1±

15.1

Sham -16.4±

9.6

4.7±

15.1

-18.55±

11.15

4.53±

15.19

-17.0±

11.0

-17.34±

9.97

28.66±

14.49

29.1±

13.9

Vel

ocity

(º/

s)

DEN 321.8±

145.8a

200.3±

91.6a

573.93±

64.71

-447.97±

100.00

-361.4±

237.8a

42.13±

122.98

-407.78±

156.76

-228.0±

256.1

REINN -313.4±

100.7

1.4±

56.2

386.26±

43.01

-297.82±

49.87

431.4±

113.6

846.71±

100.51

-912.5±

140.87

-402.3±

91.2

Sham -280.9±

127.4

-52.2±

81.8

359.45±

87.23

-359.57±

68.31

486.8±

104.1

1025.51±

198.15

-967.05±

189.79

-383.1±

220.6 a Significantly different from REINN and Sham; p < 0.05; b Significantly different from Sham only; p < 0.05

Table 6 - Mean values for hip joint angle and angular velocity for each experimental group; eight animals per group

Hip joint

Groups IC MSt Peak value

extension

Peak value

flexion TO

Peak value

extension

Peak value

flexion MSw

Ang

le (

º)

DEN 90.51

14.84b

112.50

8.70b

116.58

11.07

91.25

14.46a

109.42

13.51b

104.45

14.30b

74.77

15.54

75.38

15.71

REINN 65.44

10.20

74.76

13.11b

77.27

14.38

65.30

10.04

71.06

14.41

74.90

15.53

63.30

10.84

63.94

11.47

Sham 76.42

10.15

89.01

11.16

98.50

14.40

76.44

10.19

93.40

14.80

97.42

15.48

74.38

9.79

77.26

10.80

Vel

ocity

(º/

s)

DEN 294.67

85.80a

125.43

83.40

116.70

86.56

-372.18

168.03

-347.56

116.73a

302.19

82.24a

-471.49

155.71a

-138.63

86.07

REINN 62.52

32.64

73.92

48.24b

109.15

34.03

-70.02

68.37

-155.40

49.02

71.27

26.05

-164.38

64.76

-30.01

78.72

Sham 71.42

29.92

157.03

44.71

163.95

32.73

-89.67

82.09

-236.47

55.28

86.56

18.40

-299.45

62.17

-144.23

107.46

aSignificantly different from REINN and Sham; p < 0.05; bSignificantly different from Sham only; p < 0.05

Table 7 - Mean values for knee joint angle and angular velocity for each experimental group; eight animals per group

Knee joint

Groups IC MSt Peak value

extension

Peak value

flexion TO

Peak value

extension

Peak value

flexion MSw

Ang

le (

º)

DEN 121.08

14.46

105.33

16.04a

123.34

14.88

77.33

7.83

79.40

13.55

120.07

14.91

62.24

14.10b

72.13

17.87

REINN 105.54

14.30

69.66

13.70

105.92

14.71

56.29

15.09c

50.70

13.34c

105.73

14.11

46.62

11.92

64.75

12.20

Sham 119.69

11.78

80.06

7.41

120.11

11.90

72.50

11.26

68.25

12.65

119.78

11.86

59.86

10.39

76.72

10.27

Vel

ocity

(º/

s)

DEN 26.49

69.29

-161.26

84.53a

-40.14

64.96

-270.98

44.66

-621.61

137.42a

959.80

189.25

-452.17

144.97

872.07

230.25a

REINN 9.01

82.35

-255.12

28.54

-125.40

77.40

-390.89

56.54

-265.91

65.89

1101.70

105.38

-237.73

66.86

1073.30

101.82

Sham -58.44

54.69

-213.98

69.05

17.96

112.34

-447.77

84.79

-363.16

87.63

1173.84

139.00

-339.01

110.13

1100.27

173.29

a Significantly different from REINN and Sham; p < 0.05; b Significantly different from REINN only; p < 0.05 c significantly different from DEN and

Sham; p < 0.05




4.3 CYCLOGRAMS

The cyclograms in Figure 31 display the interjoint coordination for hip-knee, hip-ankle

and knee-ankle during the rat walking cycle. Our cyclograms of Sham group are

comparable to thoe observed in non-injured rats during level walking (29).

In line with hip motion changes, hip-knee plot in the DEN group displays wider

perimeter when compared to the REINN group (p=0.000). This increased perimeter

results from enlarged hip joint range of motion during walking in DEN group that is

approximately twice this joint’s range of motion in the other groups. Aside the higher

perimeter, the hip-knee plot also reveals interjoint coordination changes in the DEN

group. The stance-to-swing transition in sham controls (Sham group) is characterized

by reversal of hip motion from extension to flexion with the knee keeping its flexion

position during the early part of the swing phase. In contrast, hip flexion in the DEN

group initiates with the rat’s foot still in ground contact and fulfils approximately half of

its flexion range of motion before the foot leaves the ground (swing initiation). Similarly,

hip extension in the DEN group starts well in the swing phase, with the foot still off-

ground and the knee at its maximal extension angle. Hip-knee plot perimeter is similar

in REINN and Sham groups as well as the shape of the angle-angle contour. However,

in the REINN group the hip-knee loop is markedly shifted to the left, reflecting the fact

that the hip joint is operating at a more flexed position.


120 Functional assessment after Peripheral Nerve Injury: Kinematic Model of the Hindlimb of the rat

The cyclogram for the hip-ankle displays a horizontal eight-shaped pattern, present

also in the DEN group. In the REINN and Sham groups, the ankle joint at IC joint

reaches maximal plantarflexion and the hip has initiated extension. The DEN group,

however, shows altered changes in hip-ankle pattern at this time point, with the ankle

Figure 31 Ankle-knee, knee-hip and ankle-hip loops during the walking cycle for DEN (single line),

REINN (triple line) and Sham (double line) groups. Full lines correspond to the stance phase and

dashed lines to the swing phase. The start of the walking was taken as the initial paw ground

contact and is marked by a full circle. The end of the swing phase is signed with an arrowhead.




moving already to dorsiflexion and with the hip having completed a significant amount

of its extension range of motion. Again, the major distinction between REINN and

Sham groups is higher hip flexion in the REINN group that is accompanied by slightly

increased ankle plantarflexion. During the stance phase, the range of ankle dorsiflexion

in the DEN group is increased when compared to both REINN and Sham groups, as

described when presenting individual joint kinematics. The hip-ankle loop pertaining to

the DEN group demonstrates that ankle plantarflexion during the stance phase is

associated with hip flexion. This suggests that flexion of the hip joint is a strategy that

enables foot clearance at the end of stance in sciatic-denervated animals, and

compensates the lack of contractile force by plantarflexor muscles.

In the REINN and Sham groups, TO coincide with maximal ankle plantarflexion

whereas the knee is maximally flexed at this walking event (see knee-ankle plot, Fig.

2). In these groups, ankle performs a fast dorsiflexion movement during the first half of

the swing phase while the knee maintains static its flexion angle. Ankle plantarflexion

and knee extension occur during the second half of the swing phase, reaching maximal

extension just before IC. In the DEN group, the major change in knee-ankle

coordination occurs during the swing phase and is characterized by active knee flexion

early during the swing phase, most likely to help raise the paralysed foot off the ground.

The perimeters of ankle-knee cyclograms were similar in all groups. Knee-hip

cyclogram perimeter shows differences between DEN and REINN groups. Hip-ankle

cyclogram perimeter is significantly different in DEN group compared to the REINN and

SHAM groups (p= 0.001 and p=0.000 respectively).

4.4 Discriminant analysis

Four different discriminant models were built: one model based on spatiotemporal data

and three other models based on the kinematic data of hip, knee and ankle joints. Peak

angle and angular velocity values were not considered for model development. All

models correctly classified DEN group animals, excepting the model constructed on the

basis of spatiotemporal data that assigned one of animal of the DEN group into the

REINN group. Models based on kinematics of the hip, knee and ankle joints showed

full sensitivity to identify the animals of the DEN group. Also, none of the animals of

REINN and Sham groups was wrongly classified into the DEN group by any of the

models. Therefore, the kinematics of the hip, knee or ankle all provide maximal

specificity for detecting walking changes in the early phase after sciatic crush injury.



However, the several discriminant models revealed significantly less ability to identify

animals of the REINN and Sham groups. In REINN group, the rate of correct

assignments, i.e. sensitivity, lowered to 50% when using spatiotemporal data.

Surprisingly, the best models to distinguish between animals of the REINN and Sham

groups were those based on hip and knee kinematics. Assuming sensitivity as the

ability to identify animals of the REINN group, both hip and knee kinematics models

reach a sensitivity of 87.5% (7 out of 8 animals correctly classified), remaining

unchanged with cross-validation. However, the hip kinematics-based model displayed

better specificity, misclassifying just two animals of the Sham group before cross-

validation, and three after cross-validation. In contrast, the discriminant model based

upon the ankle joint kinematics performed poorly, correctly classifying just five animals

(62.5%) of the REINN group. The least sensitive model in finding the sciatic-

reinnervated animals was that employing spatiotemparal data. The hip kinematic

parameters selected by the stepwise method of linear disciminant model building were

hip angle at TO and MSw and hip angular velocity at IC and MSw.




5 Discussion

The assessment of joint kinematics during walking can be useful to evaluate functional

recovery after hindlimb peripheral nerve injury in the rat only if its sensitivity and

specificity are acceptable. Here, we tested a large set of spatiotemporal and joint

kinematics parameters describing rats’ walking function, taking for the first time into

account data from the hip, knee and ankle joints. We showed that the large majority of

spatiotemporal and joint parameters are affected early post sciatic nerve crush injury.

Moreover, we demonstrated the ability of joint kinematics to highlight minor walking

changes after long term recovery from sciatic crush injury. However, the kinematics

changes during walking twelve weeks after sciatic nerve crush were noticed in the hip

and knee joints, while ankle joint kinematics had recovered its normal pattern. In line

with joint kinematics data, discriminant analysis clearly demonstrated that walking

measures, either pertaining spatiotemporal features or hindlimb joint kinematics

measures, accurately recognize sciatic-denervated animals and distinguishes them

from sciatic-reinnervated animals and sham controls. Only the models employing hip

and knee kinematics demonstrated good accuracy when distinguishing animals that

had recovered from sciatic crush from sham controls.

5.1 Spatiotemporal and joint kinematics in sciatic-denervated

and reinnervated animals

Different factors affect the accuracy and variability of hindlimb joint kinematics during

the rat walking, including two- versus three-dimensional analysis (30), the use of

anatomical- versus reference-based system (29), skin slippage and the difficulty in

tracking bony references (31), over ground versus treadmill walking (32),

familiarization, walking velocity and the size and weight of animals (14). However,

despite the various sources of variation ankle joint kinematics during walking in the rat

displays good day-to-day and inter-observer reliabilities when assessed by 2D video

recordings (14). In our study, we adopted a very careful experimental design in order to

avoid many of the above mentioned sources of potential bias and to reach accurate

measures of hindlimb joint kinematics during walking in rats with different degrees of

deficits and in control animals. In this sense, we used a motion capture system with

high frame rate and notable tracking precision, thus avoiding data aliasing and spatial



imprecision. Great care was also taken in placing the reflective markers, and this task

was always carried out by the same person that followed a strict procedure. In addition,

the number of animals in each group was relatively large, the age and weight of sham-

operated and REINN animals were similar, and all animals were familiarized with

walking in the corridor prior to data collection. Therefore, our data can be viewed as a

reference for sagittal hip, knee and ankle kinematics in rats after sciatic nerve crush.

In the first week post sciatic nerve crush, and in the rat, the regenerating axons are

elongating in direction of the target organs but the end plates of the target muscles

remain denervated (23), which produces paralysis of denervated muscles of the

hindleg and foot and severe walking deficits. However, the animals are still capable of

locomotion by adopting an abnormal walking pattern that affects either the

spatiotemporal features of gait, the joint kinematics, and that is recognizable by simple

observation. In the DEN group, we observed slower walking speed and shortened

stride length. Also in this group, the stride duration was increased mainly caused by a

longer swing phase, while the stance phase duration seemed not to be affected. In

contrast, the gait spatiotemporal characteristics in the REINN group were similar to

those of the Sham group, indicating total recover of these gait parameters in twelve

weeks post sciatic crush, which is in line with previous reports (33). In fact, the

recovery of the spatial and temporal features of gait to pre-injury levels takes around

four weeks in cases of sciatic nerve crush in the rat (33). The mean walking speed in

our study varied from around 16 cm/s, in the DEN group, to 23 cm/s in sham controls,

with intermediate velocity of 21 cm/s in the REINN group. Walking velocity in rats

varies in a wide range and velocities between 20 to 50 cm/s are considered low (34).

This signifies that our reported velocities are at the lower limit of normal rat walking

velocity, even in our sham-controls. However, we were very careful selecting the

walking cycles for analysis and only those from uninterrupted sequences of three to

five walking steps were chosen, discarding steps with arrests or accompanied by

evident exploratory behaviours like sniffing.

Ankle joint showed the largest motion changes during walking early post sciatic injury,

comparing to the hip and knee joints. This is corroborated by several previous studies

using comparable sciatic injuries (5; 14; 25). Most of these studies have focused on the

stance phase of walking and they generally report increased ankle dorsiflexion in the

weeks immediately after the nerve injury (5; 7; 11; 12; 16; 17). Varejao et al. (19), using

computerized high-speed (225 images per second) video system, show increased

ankle dorsiflexion (90º angle between leg and foot corresponds to neutral ankle

position) during the entire stance phase when animals are assessed one week after




sciatic nerve axonotmesis. Moreover, the degree of ankle dorsiflexion rises from the

point of initial contact to near the end of the stance phase. Although we employed a

photogrammetric system, our results are in total agreement with this latter study

(Figure 30). However, ankle motion changes in the DEN group during the swing phase

were also substantial. This is not a unique observation and decreased ankle

dorsiflexion angle at midswing, defined as the instant where the swinging leg crosses

the contralateral hindlimb, has already been reported after sciatic nerve crush (14; 18).

Our data confirms such decreased dorsiflexion at MSw plus decreased plantarflexion

angle in the late swing phase, which is compatible with lack of contractile ability by the

dorsiflexor and plantarflexor muscles supplied by the peroneal and tibial branches of

the sciatic nerve, respectively. Moreover, looking at both ankle angle and angular

velocity trajectories (Figure 30) it emerges that it is in the swing phase that ankle

motion pattern in the DEN group most deviates from the REINN group and the Sham

group. However, since we are not comparing the whole trajectories but simply testing

differences in certain kinematic parameters drawn from these trajectories at specific

time points, we found similar ankle angular velocity at MSw between all groups. This

apparently paradoxical finding is enlightening and shows how is critical the choice of

the individual kinematics parameters and exemplifies the chance of finding unaffected

parameters in animals presenting severe walking deficits (11).

Our data shows complete recovery of the normal ankle kinematics during walking in

twelve weeks after the sciatic crush. This agrees with prior studies with comparable

sciatic nerve injury, showing total recovery of ankle motion during walking in 2 to 6

weeks, depending on the particular angle parameter (5; 11; 14; 18). Notwithstanding

the fast recovery in ankle joint kinematics, we and other authors have reported subtle

and persisting ankle joint motion changes, emerging later in the recovery period (11;

12; 14). However, the exact meaning of these ankle motion changes is unclear. They

might indicate abnormal reinnervation of target organs and axon misrouting, causing

diskynesis that manifests only in later stages of recovery (18; 35); but the hypothesis

that in the sciatic crush model they simply reflect the effect of the various sources of

variation affecting the precision of these measures is also a strong possibility,

particularly taking into account their inconsistency in repeated measures taken at

regular intervals (11; 18).

A major contribution of this study is the reported changes in hip and knee kinematics

during walking affecting sciatic-injured rats. This improves our knowledge about the

extent of the walking deficits caused by sciatic injury but is particularly relevant since

altered patterns of hip and knee motion were encountered not just in the DEN group



but also in the REINN group. In the DEN group, hip and knee motion changes were

expected as a mean to overcome the dysfunction at the ankle and foot and for allowing

locomotion. Higher level tasks required for walking include foot clearance from the

ground and the ability to protract and retract the whole limb during the swing and

stance phases, respectively (36). The ankle joint actively contributes to these tasks

either by joint rotation, particularly important in the swing phase, or by providing joint

stability through contraction of the muscles actuating this joint, which is required for

load bearing control during the stance phase. We observed increased hip and knee

extension angle in the DEN group during the stance phase. This change in normal hip

and knee action are most likely a strategy that compensates the inability of the ankle to

perform active plantarflexion. The higher hip range of motion in the DEN group is likely

a mean to replace the ankle joint in its functions of both push off and paw forward

displacement during the stance and swing phases, respectively. The hindlimb

protraction seems to be assisted also by trunk flexion, which is indicated by a visible

rounded posture of the rat’s torso. However, the contribution of the trunk to locomotion

and to the displacement of the affected hindlimb was not objectively measured. In turn,

increased knee extension during the stance phase maintains limb length in the

absence of ankle plantarflexion (Table 6).

Our data suggests that the hip plays a very important role in keeping limb function

during locomotion. In each phase of the walking cycle, the hip undergoes a steady

flexion or extension motion. This steady motion is lost in denervated animals. In the

DEN group, hip flexion starts before toe off in the stance phase, whereas hip extension

begins with the paw still off ground during the swing phase. This changed pattern of hip

motion probably emerges as a solution to move the paralytic hindleg using mechanical

energy transfer from the hip up to the foot. In the REINN group, the range of motion of

the hip joint during walking is similar to sham controls but the joint operates in

increased flexion, as demonstrated by the leftward shift of the loop in both the knee-hip

and ankle-hip plots of the REINN group (Figure 31). This change in hip motion was

unsuspected and could not be detected by simple observation. The reasons explaining

such altered hip motion in the REINN group cannot be resolved by our study. However,

one simple hypothesis would consider the increased hip flexion as an after effect of the

coordination pattern developed at early stages of severe deficit. As we saw,

denervated animals rely on trunk and hip flexion to propel the affected limb and this

facilitation of flexion may not be resolved even after recovery of the normal ankle joint

motion. Other possibility looks at the hip motion changes as an adaptive strategy

implemented at whole limb level to adjust for deficits at the ankle that cannot be




disclosed by biomechanical analysis of this individual joint. In the cat, triceps surae and

gastrocnemius self-reinnervation are at the origin of motion deficits that surpasses the

ankle joint motion and affect the kinematics of the other limb joints (36). Apparently,

there are mechanisms that adjust individual joint kinematics in a step-by-step basis to

keep a relatively invariant whole limb orientation and length across the walking cycle

(36). Though, it is possible that hip motion changes during walking in our reinnervated

rats reflect the action of mechanisms that maintain a stable locomotor function in the

presence of long-lasting or permanent deficits in force generation ability or

proprioceptive feedback of the reinnervated muscles (1; 20). Probably more stringent

walking tasks, such as up-hill or downhill walking, are needed to unveil the ankle joint

motion deficits in the long term after sciatic nerve crush in the rat (20).

5.2 Discriminant analysis

Several tests based in walking analysis have played an extremely useful role in

assessing functional recovery in peripheral nerve research. To assess the efficacy of

novel peripheral nerve treatments we require functional test that are very accurate

since alternative nerve treatments will most likely be characterized by moderate or

small effect size (5; 37). The development of valid and accurate tests for unbiased

assessment of functional recovery in the rat is a challenging task considering the

difficulty in standardising the testing conditions and the constraint imposed by reduced

number of subjects in the experimental groups. Ankle joint kinematics during walking is

envisaged as a reliable and accurate test to assess function recovery but we should

underline that its testing properties have been evaluated using gross functional deficits

and contrasting groups of animals with deficits with large differences in magnitude (15;

16; 24). Here, we directly addressed this limitation by comparing sham controls with

animals that had recovered from sciatic crush injury for twelve weeks, a recovery time

that is enough to re-establish the normal values in many functional tests (38; 39).

Corroborating the data from several of the measured parameters, our discriminant

analysis confirmed the sensitivity of ankle kinematics during walking to highlight

functional deficits in sciatic-denervated animals. This level of sensitivity, however, is

shared by the discriminant models developed on the basis of hip and knee kinematics

as well as on spatiotemporal measures of the rat walking. This means that test

sensitivity is not a critical point when testing animals with deficits of great magnitude

caused by peripheral nerve injury, and researchers are comfortable choosing from a

rather large set of tests to evaluate functional deficits (7; 19; 25; 32; 33; 39; 40). The

critical issue is, therefore, how can we measures small differences in functional



recovery that might indicate enhanced nerve regeneration, better re-establishment of

end organ reinnervation or changes at the integrative neural circuits at spinal or

supraspinal levels using behavioural tests. To date, we were unable to find evidence in

the literature that current tests employed to assess functional recovery after sciatic

nerve injury demonstrate such potential, particularly considering commonly employed

static tests (33). In an attempt to help solving this drawback, we explored the potential

of discriminant analysis. This statistical method is a powerful and popular classifier

technique and has been previously applied to extract predicting features from complex

movement patterns (41). Surprisingly, the discriminant models that employed hip and

knee joints parameters showed a very good sensitivity in distinguishing between

animals of the REINN and Sham groups. In fact, both hip- and knee-based discriminant

models identified correctly seven out of eight animals belonging to the REINN group.

This is a good sensitivity level and largely surpassed the sensitivity of the models

based on spatiotemporal or ankle joint parameters. Of note, the level of agreement

reached by the hip- and knee-based discriminant models was robust and did not

diminished with cross-validation.

Many of the variables selected by the discriminant analysis stepwise procedure regard

joints’ angular velocity. In the case of hip-based discriminant model, the predictor

variables selected included hip angular velocities at IC and MSw plus hip angles at TO

and MSw. Unlike joint angles, angular velocity is affected by gait speed and this might

be a limitation in using angular velocity data when animals walk freely along the

corridor. Although we could not find significant differences in group means for

spatiotemporal data between the REINN and DEN groups, we found significant

correlations between walking velocity and some of the angular velocity parameters

when considering only data from these two groups (data not shown). It is possible then

that differences between groups in angular velocity data reflect our particular

experimental setup and self-paced locomotion. However, our discriminant model using

spatiotemporal data consistently showed poorer performance when compared with

models using joint kinematics data. Further studies may want to look whether our

results are reproduced in more constrained walking conditions, such as using a

motorized treadmill.




6 Conclusions

Hip, knee and ankle joint kinematics during walking are all deeply affected following

sciatic nerve crush in the rat. After twelve weeks, animals recover the normal pattern of

ankle joint motion but kinematics changes affecting the hip and knee joints subsist. The

discriminant models using kinematics data of the ankle, knee and hip joints reveal

maximal sensitivity and specificity to identify the denervated rats against reinnervated

and control counterparts. However, poor classification performance characterized most

of the discriminant models when addressing reinnervated and control animals. From all

tested spatiotemporal and joint kinematics variables, those describing hip joint motion

revealed higher ability to highlight minor walking deficits in the reinnervated animals.

We conclude that hip and knee kinematics analysis improves the sensitivity of rat

walking test and its potential to single out subtle functional changes in the long term

after sciatic injury, which may be essential to prove the advantage of new treatment

approaches.



References

1. English AW, Chen Y, Carp JS, Wolpaw JR, Chen XY. Recovery of

electromyographic activity after transection and surgical repair of the rat sciatic

nerve. [Internet]. Journal of neurophysiology. 2007 Feb;97(2):1127-34.

2. Brushart TM. Preferential reinnervation of motor nerves by regenerating motor

axons. [Internet]. The Journal of neuroscience : the official journal of the Society

for Neuroscience. 1988 Mar;8(3):1026-31.



[Internet]. Brain research reviews. 2006 Sep;52(2):381-8.


Chapter 11: Tissue engineering of peripheral nerves [Internet]. Elsevier; 2009. 0



regeneration in an axonotmesis rat model. [Internet]. Biomaterials. 2008

Nov;29(33):4409-19.



reconstruction of the rat sciatic nerve [Internet]. Microsurgery. 2007;27(2):125–

137.



standardized crush injury in the rat. [Internet]. Microsurgery. 2008 Jan;28(6):458-

70.

8. Brushart TM, Jari R, Verge V, Rohde C, Gordon T. Electrical stimulation restores

the specificity of sensory axon regeneration. [Internet]. Experimental neurology.

2005 Jul;194(1):221-9.

9. English AW, Cucoranu D, Mulligan A, Sabatier M. Treadmill training enhances

axon regeneration in injured mouse peripheral nerves without increased loss of

topographic specificity. [Internet]. The Journal of comparative neurology. 2009

Nov ;517(2):245-55.

10. Angelov DN, Ceynowa M, Guntinas-Lichius O, Streppel M, Grosheva M,

Kiryakova SI, et al. Mechanical stimulation of paralyzed vibrissal muscles

following facial nerve injury in adult rat promotes full recovery of whisking.

[Internet]. Neurobiology of disease. 2007 Apr ;26(1):229-42.







methods. 2007 Jun ;163(1):92-104.



nerve. 2002 Nov;26(5):630-5.




2004 Nov;21(11):1652-70.

14. Ruiter GC de, Spinner RJ, Alaid AO, Koch AJ, Wang H, Malessy MJ a, et al.

Two-dimensional digital video ankle motion analysis for assessment of function

in the rat sciatic nerve model. [Internet]. Journal of the peripheral nervous

system : JPNS. 2007 Sep;12(3):216-22.



16. Lin FM, Pan YC, Hom C, Sabbahi M, Shenaq S. Ankle stance angle: a functional

index for the evaluation of sciatic nerve recovery after complete transection.

[Internet]. Journal of reconstructive microsurgery. 1996 Apr ;12(3):173-7.

17. Amado S, Rodrigues JM, Luís AL, Armada-da-Silva P a S, Vieira M, Gartner A,

et al. Effects of collagen membranes enriched with in vitro-differentiated N1E-

115 cells on rat sciatic nerve regeneration after end-to-end repair. [Internet].

Journal of neuroengineering and rehabilitation. 2010 Jan;77.



in the rat sciatic nerve model. [Internet]. Experimental neurology. 2008

Jun;211(2):339-50.



[Internet]. Muscle & nerve. 2003;27(6):706–714.

20. Maas H, Prilutsky BI, Nichols TR, Gregor RJ. The effects of self-reinnervation of

cat medial and lateral gastrocnemius muscles on hindlimb kinematics in slope

walking. [Internet]. Experimental brain research. 2007 Aug;181(2):377-93.



21. Sefton JM, Hicks-Little CA, Hubbard TJ, Clemens MG, Yengo CM, Koceja DM,

et al. Sensorimotor function as a predictor of chronic ankle instability. [Internet].

Clinical biomechanics (Bristol, Avon). 2009 Jun;24(5):451-8.

22. Press SJ, Wilson S. Choosing Between Logistic Regression and Discriminant

Analysis [Internet]. 2007 Nov ;

23. Jaweed MM, Herbison GJ, Ditunno JF. Denervation and reinnervation of fast

and slow muscles. A histochemical study in rats. [Internet]. The journal of

histochemistry and cytochemistry : official journal of the Histochemistry Society.

1975 Nov ;23(11):808-27.














;47(1):31-9.

29. João F, Amado S, Veloso A, Armada-da-silva P, Maurício AC. Anatomical

Reference Frame versus Planar Analysis : Implications for the Kinematics of the

Rat Hindlimb during Locomotion. Reviews in the Neurosciences. 2010 ;485469-

485.

30. Couto P a, Filipe VM, Magalhães LG, Pereira JE, Costa LM, Melo-Pinto P, et al.

A comparison of two-dimensional and three-dimensional techniques for the

determination of hindlimb kinematics during treadmill locomotion in rats following

spinal cord injury. [Internet]. Journal of neuroscience methods. 2008 Aug

;173(2):193-200.

31. Filipe VM, Pereira JE, Costa LM, Maurício AC, Couto P a, Melo-Pinto P, et al.

Effect of skin movement on the analysis of hindlimb kinematics during treadmill

locomotion in rats. [Internet]. Journal of neuroscience methods. 2006 May

;153(1):55-61.




32. Pereira JE, Cabrita AM, Filipe VM, Bulas-Cruz J, Couto P a, Melo-Pinto P, et al.

A comparison analysis of hindlimb kinematics during overground and treadmill

locomotion in rats. [Internet]. Behavioural brain research. 2006 Sep ;172(2):212-

8.

33. Bozkurt A, Deumens R, Scheffel J, OʼDey DM, Weis J, Joosten EA, et al.

CatWalk gait analysis in assessment of functional recovery after sciatic nerve

injury. [Internet]. Journal of neuroscience methods. 2008 Aug ;173(1):91-8.

34. Koopmans GC, Deumens R, Brook G, Gerver J, Honig WMM, Hamers FPT, et

al. Strain and locomotor speed affect over-ground locomotion in intact rats.

[Internet]. Physiology & behavior. 2007 Dec ;92(5):993-1001.

35. Ijkema-Paassen J, Meek MF, Gramsbergen A. Reinnervation of muscles after

transection of the sciatic nerve in adult rats. [Internet]. Muscle & nerve. 2002 Jun

;25(6):891-7.

36. Chang Y-H, Auyang AG, Scholz JP, Nichols TR. Whole limb kinematics are

preferentially conserved over individual joint kinematics after peripheral nerve

injury. [Internet]. The Journal of experimental biology. 2009 Nov ;212(Pt

21):3511-21.

37. Yao L, Ruiter GCW de, Wang H, Knight AM, Spinner RJ, Yaszemski MJ, et al.

Controlling dispersion of axonal regeneration using a multichannel collagen

nerve conduit. [Internet]. Biomaterials. 2010 Aug ;31(22):5789-97.

38. Bodine-Fowler SC, Meyer RS, Moskovitz A, Abrams R, Botte MJ. Inaccurate

projection of rat soleus motoneurons: a comparison of nerve repair techniques.

[Internet]. Muscle & nerve. 1997 Jan ;20(1):29-37.

39. Hare GM, Evans PJ, Mackinnon SE, Best TJ, Bain JR, Szalai JP, et al. Walking

track analysis: a long-term assessment of peripheral nerve recovery. [Internet].

Plastic and reconstructive surgery. 1992 Feb ;89(2):251-8.

40. Varejão a S, Meek MF, Ferreira a J, Patrício J a, Cabrita a M. Functional

evaluation of peripheral nerve regeneration in the rat: walking track analysis.

[Internet]. Journal of neuroscience methods. 2001 Jul ;108(1):1-9.

41. Cavalheiro GL, Almeida MFS, Pereira AA, Andrade AO. Study of age-related

changes in postural control during quiet standing through linear discriminant

analysis. [Internet]. Biomedical engineering online. 2009 Jan ;835.

Chapter 6 - The effect of active and passive exercise in

nerve regeneration and functional recovery after sciatic

nerve crush in the rat

S Amado, AL Luis, S Raimondo, M Fornaro, MG Giacobini-Robecchi, S Geuna, F

João, AP Veloso, AC Maurício , PAS Armada-da-Silva (2012). Neuroscience

(submitted)



1. Abstract

For the first time the effects of exercise after peripheral nerve injury was assessed with

quantitative method of kinematic analysis. There is growing evidence that both active

and passive exercise improve peripheral nerve regeneration and functional recovery.

With the aim of contributing to this mounting evidence, we investigated the effect of

treadmill walking exercise and the effect of passive mobilization of the entire hindlimb

in the sciatic morphology and functional recovery after a sciatic crush injury in the rat.

The injured animals were separated in a treadmill exercise group (EXa), passive

mobilization group (EXp) and a group without any further treatment intervention

(Crush). A sham-operated control group was also included (Sham). Active exercise

consisted on 30 minutes daily of treadmill walking at a speed of 10 m/min, while

passive mobilization consisted on 3 minutes repetitive triple flexion and triple extension

(i.e. simultaneous flexion or extension of the hip, knee and ankle joints) with slight

stretching at the two movement extremes. The two types of exercise slightly improved

nerve morphology at 12-weeks survival time, reducing the total number of regenerated

myelinated fibers to values similar to intact nerves. This effect of exercise on nerve

regeneration seems to operate through different mechanisms dependent on whether

the exercise is active or passive, since nerve fibers density was significantly increased

in the Exp group, in contrast to the sciatic-injured groups. Functional recovery was

assessed using gait variable and hindlimb joint kinematics during level self-paced

walking. At the end of the 12-weeks survival time spatiotemporal gait variables were

similar in all groups. Similarly, ankle kinematics recovered the normal motion pattern in

all groups. In contrast, hip and knee kinematics in the Crush group were still abnormal

at the end of the survival time, while these joints in both exercise groups displayed a

pattern of motion during walking that was indistinct from that of the sham control

animals. It is concluded that either active or passive exercise positively affect sciatic

nerve regeneration after a crush injury. We suggest that the positive role of passive

exercise results from a general stimulus induced by the movement and possibly

mediated by a direct mechanical effect onto the regenerating nerve.

Chapter 6 – The effect of active and passive exercise on nerve regeneration and functional recovery after sciatic

crush injury 137


2 Introduction

Peripheral nerve injuries cause motor, sensory and autonomic dysfunction in the

denervated territory. After injury and repair, the injured axons grow from the proximal

nerve stump in direction to the target organs. However, and irrespective of the good

ability of peripheral nerves to regenerate, the degree of reinnervation and of functional

recovery after peripheral nerve injury is typically disappointing (1). Dysfunction after

peripheral nerve injury and repair is manifold. In the rat sciatic nerve crush model,

several studies show increased number of nerve fibers as well as decreased nerve

fibers’ diameter and decreased myelin thickness in regenerated nerves (2-4). The

increase in the number of nerve fibers distally to the injured site results from multiple

branching of the severed axons (5-7). The many collaterals branching from single

proximal axon often disperse and lead to inappropriate reinnervation, including

reinnervation of multiple muscle groups possibly with antagonistic function (8). A

second wave of axonal sprouting occurs when regenerating axons reach the target

muscle. In this case, axon terminal splits in several small sprouts that course through

bridges formed by extensive Schwann cells processes. In this way, multiple fine axonal

branches spread to re-innervate multiple end-plates. The poly-neuronal innervation

caused by increased density of sibling collaterals reaching the denervated muscle

groups and by terminal axon sprouting, disturbs force generation control and

contributes to poor functional recovery (8). Sciatic nerve injury also affects sensitive

function. Deafferentiation causes adaptive changes at the dorsal root ganglion and the

dorsal horn of the spinal cord, disrupting the normal processing of afferent inflow and

spinal reflexes integration (9). At a higher level, remapping of central controlling neural

circuitry may occur in order to cope with muscle paralysis and to maintain motor

function. All these changes contribute to dyskinesia, dysreflexia and hyperalgesia that

typically develop after peripheral nerve transaction and that severely compromise

functional recovery.

The poor functional outcome after peripheral nerve repair and its clinical relevance

have raised the interest in the development of rehabilitation strategies to enhance

nerve regeneration and improve functional recovery. Physical exercise is one of such

rehabilitation strategies and has been the focus of intense research recently. Many

different types of exercise have been tested, including active (3; 10-14) and passive

exercise (15; 16). Treadmill walking is a common procedure to stimulate the

denervated territories and the regenerating axons in rodents. This type of exercise



uses a natural pattern of locomotion and its intensity can be easily adjusted, turning

treadmill walking a very useful exercise model. Studies in mice show that active

treadmill walking at both mild and high intensity increases the rate of axonal elongation

after nerve transection (12) as well as it leads to a higher intensity in axonal collateral

branching but without accompanying increase in incorrect reinnervation (17). The latter

effect contrasts to that of electrical nerve stimulation delivered at the time of nerve

repair that has been shown to increase collateral axonal branching at the lesion site but

also causing excessive axonal misrouting (18). Positive effect of treadmill walking is

also reported for adult rats (16). In these animals, daily treadmill walking produced an

increase in the number of myelinated nerve fibers in transected and end-to-end

repaired sciatic nerves and enhanced muscle reinnervation, demonstrated by

increased M-wave amplitude in the gastrocnemius, soleus and plantaris muscles of the

affected hindlimb (16). In this study, active walking exercise also resulted in improved

spinal reflex activity and in a decrease in H-reflex facilitation that is manifested after

peripheral nerve injury and non-exercised animals (9; 16).

The positive effect of exercise stimulus on nerve regeneration and in the extent and

precision of target organs reinnervation seems to apply also to passive exercise.

Increased mechanical stimulation of the vibrissal muscles and whisker muscle pad of

rats restored the normal vibrissae function after transection and direct coaptation of the

facial nerve (15). Here, manual stimulation had no effect on the degree of axonal

collateral branching, which occurs in a large extent after facial nerve injury in rats, but

instead reduced the amount of motor end-plate polyinnervation (15). This positive

effect of manual mobilization of the denervated facial nerve territory seems to rely on

intact sensory function conveyed by uninjured trigeminal nerve sensory endings (19). In

the rat sciatic nerve model, passive exercise has also been shown to positively affect

nerve regeneration and muscle reinnervation, demonstrating that the beneficial effect

of passive mobilization after peripheral nerve injury also applies to lesions of mixed

nerve trunks and is not strictly dependent on normal sensory input (16).

By causing muscle paralysis, denervation is associated with reduced movement of the

affected limbs and mechanical unloading of muscles and other soft tissues. The lack of

normal movement pattern of the joints and surrounding tissues leads to the

development of tissue contractures and joint ankylosis in severe cases. These

conditions may be prevented by passive exercise, joint mobilization and gentle muscle

stretching. By helping maintain the normal muscle structure and joint mobility, this type

of exercise may contribute to functional recovery.

In this study we tested for the effect of brief manual mobilization of the hindlimb joints

and muscles on nerve regeneration and functional recovery after sciatic nerve crush in


crush injury 139


the rat, and compared to the effect of active treadmill walking and to non-treated

conditions. Our manual mobilization technique involved moving the hip, knee and ankle

joints through their full range of motion thus applying a slight stretching to the muscles

crossing these joints as well as to the other soft tissues.



3 Methods

A total of 32 adult Sprague-Dawley male rats were randomly separated in the following

groups: (1) sciatic crush plus treadmill-walking (EXa), (2) sciatic crush plus passive

exercise (EXp), (3) sciatic crush (Crush) and (4) sham-operated control (Sham). Our

procedure to cause sciatic nerve crush has been detailed elsewhere. Briefly, we

utilized a non-serrated clamp (Institute of Industrial Electronic and Material Sciences,

University of Technology, Vienna, Austria), that exerts a constant force of 54 N onto the

nerve. The pressure was applied for duration of 30 seconds, generating a 3mm-long

crush of the sciatic nerve and complete axonotmesis. The right sciatic nerve was

crushed 1 cm above the bifurcation into tibial and common fibular nerves. This

procedure was demonstrated to cause complete axonotmesis of the sciatic nerve (4;

20-22).

Animals in the EXa group exercised for 30 min on a 10-lane motorized treadmill at a

constant speed of 10 m/min with zero incline. The exercise was performed 5 days a

week starting at week 2 post-injury and maintained until week 12 post sciatic nerve

crush (Figure 32). At least two weeks before sciatic crush injury, animals in the EXa

were acclimated to treadmill walking during five to ten minutes daily for one week.

Twelve days after the surgery, rats were again tested to ascertain their ability to walk

on the treadmill. All animals initiated the walking training protocol at day fourteen after

sciatic nerve crush.

Figure 32 - image of rats on the treadmill after sciatic nerve injury.

Passive exercise consisted on manual mobilization of the entire right hindlimb. The

passive mobilization of the hindlimb included movement of the hip, knee and ankle

joints in concert, and alternating flexion and extension of the joints. During extension, a

slight stretching was applied at the end of the movement amplitude. The passive


crush injury 141


exercise sessions took approximately three minutes and were undertaken with animals

fully awake with the head and upper torso covered with a scarf to help keep the

animals relaxed. The animals tolerated the passive exercise sessions without signs of

pain or discomfort. The exercise was performed 5 days a week and along the same

recovery period as the EXa group.

Animals of the Crush and Sham groups remained in their cages during the 12-weeks

recovery period with no other intervention. All animals were socially housed within a

single room. Dry food and water were provided ad libitum. All procedures were

performed with the approval of the Veterinary Authorities of Portugal in accordance

with the European Communities Council Directive of November 1986 (86/609/EEC).

3.1 Kinematic analysis

Gait analysis was performed before training protocol (one week post-injury) and at

weeks 4, 8 and 12 post-injury. This analysis used an optoelectronic system of six

infrared cameras (Oqus-300, Qualisys, Sweden) operating at a frame rate of 200Hz to

record the motion of the right hindlimb during the rat level gait. After skin shaving,

seven reflective markers with 2 mm diameter were attached to the right hindlimb at

bony prominences: (1) tip of fourth finger, (2) head of fifth metatarsal, (3) lateral

malleolus, (4) lateral knee joint, (5) trochanter major, (6) anterior superior iliac spine,

and (7) ischial tuberosity. The retroreflective markers were in all instances placed by

the same operator using the marked skin as a guide. Animals walked on a Perspex

track with length, width and height of respectively 120, 12 and 15 cm. Two darkened

cages were placed at the extremities of the corridor as a mean to facilitate the animals

to cross the walking corridor. All rats previously performed two or three conditioning

trials to be familiarized with the corridor. Cameras were positioned to minimize light

reflection artifacts and to allow recording 4-5 consecutive walking cycles, defined as

the time between two consecutive initial ground contacts (IC) of the right fourth finger.

The motion capture space was calibrated regularly using a fixed set of markers and a

wand of known length (20 cm) moved across the recorded field. Calibration was

accepted when the standard deviation of wand’s length measure was below 0.4 mm.

Planar motion in the sagittal plane of the hip, knee and ankle joint was calculated with

Visual 3D software (C-Motion, Inc, Germantown, USA) by a computational procedure

implementing the dot product between the skeletal segments articulated by these

joints. Joint velocity was also calculated using the first derivative of joint angle motion.

The trajectory of the reflective markers was smoothed using a Butterworth low-pass

filter with a 6 Hz cut-off. Data for each animal at each recording session consisted on



an average of six walking cycles. Each walking cycle was time normalized by

interpolation using spline fitting. Our angles convention was: 1) decreasing hip angle

value – hip flexion; 2) decreasing knee angle value – knee flexion; 3) positive ankle

angle value – dorsiflexion. Positive angular velocity indicates increasing angle values

(i.e. hip and knee extension). This also applies to the ankle joint where a positive

angular velocity means increase dorsiflexion angle. Joint angle and angular velocity

values were measured at the instants of initial ground contact (IC), defined as the time

point where horizontal velocity of the toe reflective marker becomes zero, midstance

(MSt), the point of fifty per cent duration of the stance phase, toe-off (TO), defined as

the instant the horizontal velocity of the toe’s reflective marker increases from zero, and

midswing (MSw), defined as the point of fifty per cent duration of the swing phase.

Additional kinematics parameters recorded included peak angles and angular velocities

during both phases of the walking cycle.

Spatiotemporal parameters, including walking velocity, walking cycle duration, stride

length, stance duration, and swing duration, were directly obtain from kinematic model

developed under Visual 3D software based on horizontal displacement of the reflective

markers.

3.2 Design-based quantitative nerve morphology

At the end of the 12-weeks follow-up time, rats were anaesthetised and a 10-mm-long

segment of the sciatic nerve that included the injured portion was collected, fixed, and

prepared for morphological analysis and histomorphometry of myelinated nerve fibers.

A 10-mm segment of uninjured sciatic nerve was also withdrawn from the 8 control

animals. Immediately after collecting the nerve, rats were euthanized through an

intracardiac injection of 5% sodium pentobarbital (Eutasil®). Sciatic nerve samples were

immersed immediately in a fixation solution, containing 2.5% purified glutaraldehyde

and 0.5% saccarose in 0.1M Sorensen phosphate buffer for 6-8 hours. Specimens

were then washed in 0.1M Sorensen phosphate buffer (1.5% saccarose), post-fixed in

1% osmium tetroxide, dehydrated and embedded in Glauerts’ embedding mixture of

resins (23). Series of semi-thin transverse sections (2-µm thick) were then cut starting

from the distal nerve stump and stained by Toluidine blue for high resolution light

microscopy examination and design-based stereology (24). Systematic random

sampling and 2-D disector were adopted (25-27). The following predictors of nerve

regeneration were assessed: mean fiber density (D), total fibers number (N); circle-

fitting diameter of fiber (D) and axon (d), and the g-ratio (d/D). The sampling scheme

was designed in order to keep the coefficient of error (CE) below 0.10, a level which

assures enough accuracy for neuromorphological studies (28).


crush injury 143


4 Statistical analysis

Differences between groups for each of the variables were tested by univariate ANOVA

and the post hoc Tukey’s HSD test. Significance was accepted at p<0.05. Data are

presented as mean and standard deviation (SD). Statistical analysis was performed

using SPSS software package (version 17, SPSS Inc).



5 Results

During surgery, the sciatic nerve crushed resulted in visible crush of nerve but without

interrupting its continuity. Muscle paralysis of the operated foot was observed following

crush injury. All rats survived, with no wound infection or automutilation.

1.1 Gait spatiotemporal data

All animals could walk with steady pace in the corridor during the testing sessions

carried out in the recovery period. As a consequence of sciatic nerve injury and sham

surgery, most gait spatiotemporal features changed along the recovery time. During the

12-weeks recovery, stride length increased [F(3,84) = 17.239, p = 0.000] either in sciatic-

injured groups and Sham group [non-significant main effect of group; F(2,28) = 2.688, p =

0.066], while gait cycle time diminished [F(3,84) =19.794,p=0.000] again in all groups.

The shortened gait cycle duration was associated with decreased swing phase duration

[F(3,84) = 27.133 , p = 0.000], while that of the stance phase showed no changes in any

of the groups at any joint in time. A trend for the rise of gait speed rise with time of

recovery was noticed (p=0.054). At week 1 post-injury, gait speed ranged from

0.161±0.038 m.s-1 in EXp group to 0.116±0.028 m.s-1 in Crush group. By week twelve,

mean gait speed varied from 0.239±0.052 m.s-1 in EXa group to 0.210±0.047 m.s-1 in

Crush group (Table 8).

Table 8 - Walking velocity and cycle time data if the right hindlimb in sciatic-crushed animals and

sham-operated controls; n=8 in each group

Group

Gait velocity

(m/s)

stride lenght

(m)

cycle

duration (s)

stance duration

(s)

swing

duration (s)

W01 W12 W01 W12 W01 W12 W01 W12 W01 W12

EXa

0.15

0.04

0.24

0.05

0.14

0.01

0.17

0.01

0.92

0.16

0.73

0.17

0.23

0.05

0.20

0.03

0.20

0.05

0.13

0.01

EXp

0.16

0.03

0.23

0.09

0.13

0.01

0.16

0.02

0.85

0.16

0.76

0.20

0.24

0.09

0.25

0.06

0.17

0.03

0.13

0.01

Crush

0.12

0.03

0.21

0.05

0.13

0.01

0.15

0.02

1.11

0.27

0.73

0.10

0.28

0.04

0.26

0.03

0.19

0.03

0.12

0.01

Sham

0.15

0.03

0.23

0.04

0.14

0.01

0.16

0.01

0.96

0.13

0.70

0.09

0.28

0.08

0.25

0.04

0.17

0.01

0.13

0.01


crush injury 145


1.2 Hindlimb joint kimematics

Ankle kinematics

Sciatic nerve crush causes paralysis of the hindleg and foot muscles. Therefore, we

start by presenting ankle joint kinematics data followed by that of the hip and knee

joints.

Ankle kinematics was most affected at week 1 post sciatic nerve injury (Figure 33).

Exaggerated dorsiflexion during the whole stance phase characterizes movement

about this joint in sciatic-injured animals, with differences to sham-operated controls

being most evident at midstance and toe-off time points. Clear changes in ankle

angular velocity were also induced by sciatic injury. Ankle angular velocity at week 1

post injury shifted from negative to positive values in sciatic-injured animals either at

initial paw contact and midstance, contrasting to the negative values of sham controls.

Conversely, at toe-off time point (Table 10), ankle angular velocity was negative in

sciatic nerve-injured animals and positive in Sham group. Ankle kinematics at midswing

(Table 12) was less affected by sciatic nerve crush. At week 1, sciatic crushed animals

demonstrated a decreased ability to perform ankle dorsiflexion at midswing, compared

to animals in the Sham group, but ankle angular velocity was similar in all groups. A

good recovery of the normal pattern of ankle kinematics during walking occurred during

the 12 weeks of follow up, and no consistent differences in the extent of this recovery

were observed due to treatment modality (Figure 34).



Figure 33 - Mean trajectories for hip (upper graphs), knee (middle graphs) and ankle (lower graphs)

joint angle (left hand graphs) and angular velocity (right hand graphs) during the rat’s walking

cycle. Full lines correspond to the stance phase and dashed lines correspond to the swing phase

of the walking cycle: crush group (red line); EXa (black line); EXp (blue line); Sham Group (green

line). All groups n=8


crush injury 147


Figure 34 - Mean trajectories for hip (upper graphs), knee (middle graphs) and ankle (lower graphs)

joint angle (left hand graphs) and angular velocity (right hand graphs) during the rat’s walking

cycle. Full lines correspond to the stance phase and dashed lines correspond to the swing phase

of the walking cycle: crush group (red line); EXa (black line); EXp (blue line); Sham Group (green

line). All groups n=8.



Hip kinematics

Hip kinematics during walking was also affected as a consequence of sciatic nerve

crush. Significant differences between groups in hip kinematics were acutely observed

after sciatic denervation being these differences were still noticeable at the end of the

12-weeks recovery. In the week immediately after surgery, changes in normal hip joint

motion during walking were noticed only in angular velocity values. At this point in time,

sciatic injured animals walk with increased hip angular velocity either at initial toe

contact (i.e. swing-to-stance transition) and toe off (i.e. stance-to-swing transition),

compared to Sham group animals. Despite such changes in hip angular velocity, no

significant alterations in hip position at the relevant phases of gait were recorded in

sciatic denervated animals. Sciatic nerve reinnervation led to the recovery of the

normal hip motion pattern during walking in exercised animals (i.e. EXa and EXp

groups) but differences between Crush and Sham groups emerged. In fact, all

kinematic parameters obtained from the stance phase were significantly different in

Crush group, compared to Sham control at the 12-weeks time point. The changes were

characterized by more acute hip angles at the entire stance phase and significantly

lower hip angular velocity in Crush group.

Knee kinematics

Sciatic denervation was also accompanied by compensatory changes in knee joint

kinematics as a mean to enable successful walking. Early after sciatic nerve crush,

changes in knee kinematics occurred mainly during the stance phase. At the instant of

paw touch down, knee angle shows similar values in all groups but its angular velocity

was distinctly higher in sciatic-injured animals. At midstance (Table 11), and at week 1

post injury, sciatic denervated animals display a more extended knee, when compared

to the sham operated animals. This can be reasonably interpreted as a compensatory

response to the more plantigrade pattern of walking resulting from paralysis of the

muscle crossing the ankle joint thus avoiding excessive shortening of the overall limb

length. At this particular instant of the step cycle, knee joint velocity was not affected by

the sciatic nerve injury. Conversely, at the stance-to-swing transition knee angular

velocity in the sciatic-injured animals is greatly augmented compared to Sham group,

meaning fast knee flexion motion in the former animals. A recovery of the normal knee

pattern of motion during walking occurred in the weeks after sciatic nerve crush. This

recovery was, however, enhanced by both walking exercise and manual mobilization.

At week 12 post injury, and when compared to Sham group, animals in Crush group

walked with exaggerated knee flexion at the instants of initial toe ground contact and

toe off.


crush injury 149


Table 9 - Values of hindlimb joints joint angle and angular velocity at initial contact (IC)

Joint Group Angle (º) Velocity (º/s)

week01 week12 week01 week12

Ankle

EXa -6.64.4 -13.32.9 446.2123.4 -67.2114.6

EXp -0.64.6 -22.73.5 321.8145.8 -243.0165.0

Crush 1.04.2 -20.14.4 365.4122.8 -313.4100.7

Sham -11.74.8 -16.43.4 57.2100.8 -280.9127.4

Knee

EXa 117.1 18.42 116.17.3 -59.899.5 -153.491.5

EXp 121.114.5 118.29.6 26.569.3 -46.9134.3

Crush 115.015.6 105.514.3 23.767.2 9.082.3

Sham 104.915.8 119.711.8 -214.3106.4 -58.454.7

Hip

EXa 89.111.3 78.311.4 201.938.4 75.636.5

EXp 90.514.8 72.67.0 294.785.8 70.330.4

Crush 83.214.6 65.410.2 218.394.7 62.532.6

Sham 83.69.4 76.410.1 109.552.5 71.429.9

Table 10 - Values of hindlimb joints angle and angular velocity at toe-off (TO)

Joint Group Angle (º) Velocity (º/s)


Ankle

EXa 12.0914.32 -19.259.92 -367.97254.35 320.36109.10

EXp 17.099.43 -15.009.44 -361.44237.79 363.65146.06

Crush 23.6311.57 -14.5913.62 -447.27212.76 431.40113.60

Sham -17.899.48# -17.0311.00 74.36187.84 486.84104.15

Knee

EXa 72.3012.80 66.2910.31 -563.86183.40 -281.7283.49

EXp 79.4013.55 63.8212.44 -621.61137.42 -325.0269.60

Crush 65.5015.47 50.7013.34 -557.37108.77 -265.9165.89

Sham 67.0416.21 68.2512.65 -205.83102.05 -363.1687.63

Hip

EXa 103.328.86 93.0219.35 -336.71167.08 -173.1889.81

EXp 109.4213.51 85.079.72 -347.56116.73 -208.1267.99

Crush 93.8219.59 71.0614.41 -367.31138.26 -155.4049.02

Sham 97.4216.27 93.4014.80 -155.1076.60 -236.4755.28



Table 11 - Values of hindlimb joints angle and angular velocity at midstance (MSt)

Joint

Group

MSt angle (º) MSt velocity (º/s)


Ankle EXa 42.1010.54 11.6512.28 175.98119.89+ -134.3584.06

EXp 46.228.43 6.9912.28 200.3291.63- 23.1685.33

Crush 48.0412.96 4.2414.12 163.38118.21- 1.4356.25

Sham 13.6113.19 4.7315.17 -58.6361.23 -52.1981.85

Knee EXa 94.0515.01 75.156.85 -137.3676.77 -224.3660.35

EXp 105.3316.04 80.4214.86 -161.2684.53 -247.4135.48

Crush 91.0910.28 69.6613.70 -142.4981.28 -255.1228.54

Sham 68.3115.44 80.067.41 -121.7943.65 -213.9869.05

Hip EXa 106.2110.76 89.8016.38 109.0773.71 137.1894.40

EXp 112.508.70 84.307.20 125.4383.40 124.8944.64

Crush 99.0117.72 74.7613.11 115.2167.04 73.9248.24

Sham 93.3811.15 89.0111.16 95.1970.18 157.0344.71

Table 12 - Values of hindlimb joints angle and angular velocity at midswing (MSw)

Joint

Groups

MSw Angle (º) MSw velocity (º/s)


Ankle EXa 4.5410.39 27.3812.18 -274.44310.80 -276.94129.49

EXp 11.4011.32 22.066.99 -227.98256.05 -422.19133.67

Crush 16.0911.00 23.0915.11 -313.59292.90 -402.2891.21

Sham 35.4111.38 29.0613.91 -282.91263.38 -383.13220.56

Knee EXa 71.1819.46 73.745.71 860.40333.73 1035.40145.63

EXp 72.1317.87 74.7912.39 872.07230.25 1099.1594.48

Crush 61.8814.04 64.7512.20 976.08249.34 1073.30101.82

Sham 65.9614.70 76.7210.27 844.78165.17 1100.27173.29

Hip EXa 71.259.53 78.3812.15 109.07100.53 137.1896.65

EXp 75.3815.71 71.897.19 125.4386.07 124.8975.40

Crush 63.6913.66 63.9411.47 115.21140.41 73.9278.72

Sham 80.6311.15 77.2610.80 95.1983.82 157.03107.46


crush injury 151


1.3 Nerve stereology

Figure 35 - Representative micrographs toluidine blue stained semithin sections from EXa (A), Exp

(B), Ctrl (C) and Sham (D) groups. The presence smaller myelinated nerve fibers can be

appreciated in all axonotmesis groups (A-C) in comparison to controls (D). Morphometrical

changes have been quantified by stereology and results are reported in table 6. Original

magnification = 1,000x.

Figure 35 shows representative images of myelinated nerve fibers from nerves

belonging to the four experimental groups included in this study. The results of the

stereological analysis of myelinated nerve fibers are reported in Table 13. Overall,

regenerated sciatic nerves present increased number and density of myelinated nerve

fibers in comparison to intact nerves. The regenerated nerve fibers are also

characterized by smaller diameter and lower myelin sheet thickness. Slight effects of

exercise were observed in the morphology of the regenerated sciatic nerves. When

compared to the Sham group, the increase in total number of myelinated fibers was

only significant for the Crush group (p<0.05), suggesting that both types of exercise

prevented an excessive increase in regenerating axonal collaterals. Despite this

apparently positive effect of active and passive exercise, nerve density was increased

in EXp group only, when compared to Sham control group (p<0.05). Other

morphological features characterizing the myelinated nerve fibers, i.e. diameter of



nerve fiber (D) and axon (d) and the g-ratio (D/d), were all significantly different in

regenerated sciatic nerves, compared to the intact nerves.

Table 13 - stereological analysis of myelinated nerve fibers

Group Density of

myelinated nerve

fibers (N/mm2)

Total number of

myelinated nerve

fibers (N)

d - axonal

diameter (µm)

D - nerve fiber

diameter (µm)

d/D

(g-ratio)

EXa 14786 ± 4186 9255 ± 1618 3.90 ± 0.32 5.37 ± 0.41 0.72 ± 0.02

EXp 18146 ± 4691 8974 ± 1918 3.85 ± 0.37 5.35 ± 0.44 0.71 ± 0.03

Crush 16701 ± 4251 10473 ± 1487 3.72 ± 0.31 5.22 ± 0.42 0.71 ± 0.04

Sham 11680 ± 1282 6883 ± 1222 5.35 ± 0.35 8.22 ± 0.38 0.65 ± 0.04


crush injury 153


6 Discussion

In this study we evaluated the role of two different methods of exercise on sciatic nerve

regeneration and functional recovery after a sciatic nerve crush lesion. The two

exercise modes consisted on treadmill exercise walking and passive mobilization of the

affected hindlimb. The effect of both exercise programs was evaluated in contrast to a

non-exercise group of sciatic nerve crushed animals and sham operated controls. The

results showed marginal effects of both active and passive exercise on functional

recovery, assessed by hindlimb joint kinematics during level walking. Although the

impact of exercise on functional recovery was relatively modest it was nonetheless

accompanied by improved nerve morphology, particularly by preventing excessive

increase in total number of myelinated fibers in regenerated sciatic nerves.

Physical exercise has received much interest recently due to its potential effect in

enhancing nerve regeneration and reinnervation in different models of peripheral nerve

pathology (11; 13; 15-17; 29; 30). The results of the majority of these studies point to a

small to moderate positive effect of the active exercise stimulus in nerve regeneration

after sciatic injury and repair in rodents either when used alone (16; 17; 31) or in

combination with nerve electrical stimulation (13), as well as in the degree of

reinnervation and extent of functional recovery (12; 13; 16; 31).

To date, the exact mechanisms explaining the beneficial effect of active exercise on

nerve regenerations are still not completely elucidated; although several mechanisms

have been alluded. After sciatic nerve transection and repair, walking exercise has

been shown to increase the number of myelinated nerve fibers distal to injury site (13;

16), as well as to enhance the rate of axonal elongation in the common fibular nerve of

mice (12). After being cut, elongating axons are well known to divide profusely in a

variable number of collaterals that, afterwards, may either travel in parallel along the

same nerve or take a divergent route through alternative nerve pathway leading to

inappropriate reinnervation. Increased axonal branching, although important in the

sense that it contributes with more regenerating nerve fibers to reinnervation, has a

downside consequence if accompanied by aberrant reinnervation of target organs.

Studies in mice demonstrate that nerve regeneration is considerably enhanced by

electrical stimulation applied to the proximal nerve stump for a brief period of time just

prior to nerve repair, as well as by distal nerve stump treatment with chondroitinase

ABC, an enzyme that removes the axonal growth-inhibiting glycosaminoglycans from

chondroitin sulfate proteoglycans (18). This conclusion is driven by the use of different

retrograde fluorescent tracers applied to the tibial and common fibular branches of the



sciatic nerve and consequent enumeration of labeled motoneurons at the anterior horn

of the spinal cord (18). However, the increased number of labeled motoneurons was

most pronounced two weeks after sciatic transection, lowering to almost control levels

at four weeks survival time. Importantly, the enhanced nerve regeneration, noticed two

weeks after the sciatic nerve injury, was manifested with clear loss of the normal spinal

topographical organization of tibial and common fibular motoneurons, consequent to

elongating axon collaterals entering erroneous ending nerves, which in turn might

compromise function (18). Comparable increase in the number of labeled motoneurons

of the sciatic motor nucleus also occurred at 2-week and 4-week after sciatic nerve

transection and repair when animals performed either mild-intensity continuous or high-

intensity intermittent treadmill exercise during the two recovery periods of time, with

more robust effect with the more intense treadmill exercise protocol (17). Treadmill

exercise, however, not just enhanced nerve regeneration but importantly had such

effect without simultaneously raising inappropriate target reinnervation. In addition,

treadmill exercise seems to exert a sustained nerve regeneration enhancing effect, with

the number of tibial nerve labeled motoneurons in exercised animals increasing from

week-2 to week-4 after sciatic nerve transection and repair whereas this number

remained unchanged between these two recovery times in animals that had nerve

regeneration accelerated by treatment with electrical stimulation and chondroitinase

ABC (17; 18).

Active treadmill exercise was also shown to improve nerve regeneration in the rat after

sciatic transection and direct repair (13; 16). Morphological evaluation at the level of

the tibial nerve reveals increased number of regenerated myelinated fibers in the active

exercise group, compared to the non-exercise, control group (16). Detailed

morphometry of regenerated sciatic crushed nerve in the rat again supports a positive

role of treadmill endurance exercise in nerve regeneration (31). In this latter study,

treadmill exercise started two weeks after the sciatic nerve injury and was maintained

throughout the next five weeks, with animals of the endurance exercise group walking

at 50% of their maximal exercise capacity determined pre-operatively, corresponding to

a mean walking velocity of around 9 m/min, thus comparable to that we used in the

present study. When compared to untrained control animals, the endurance-trained

group depicted better nerve regeneration, demonstrated by greater myelin thickness,

higher area fraction of myelinated nerve fibers and less of that of endoneurial

connective tissue in regenerated sciatic nerves (31).

A major aim of the present study was to test for the effect of brief passive manual

mobilization of the affected hindlimb on sciatic nerve regeneration and functional

recovery. Our data of the sciatic nerve morphology collected distally to injury site again


crush injury 155


showed signs of better regeneration as a result of passive hindlimb mobilization and

gentle stretching. As for the EXa group, and when excluding outliers from statistical

testing, mean total number of myelinated fibers in the EXp group was indistinct from

that of the Sham group. To some extent , this agrees with other studies investigating

the effect of passive exercise on sciatic nerve regeneration after transection injury and

direct cooptation (16). Udina et al. (16) tested the effect of passive bicycling in the rat

on sciatic regeneration and noticed improved sciatic regeneration to a similar degree to

that registered for active treadmill exercise. It is difficult to interpret, and thus it

deserves further research, the discrepancy between the results of this authors who

revealed an increase in the number of myelinated axons and our results, which

evidenced the opposite.

In the case of the facial nerve, manual mobilization of the whisker pad reduced muscle

fibers poly-innervation two months after nerve transection and repair. In this case,

however, the passive exercise showed no effect on the degree of regenerating nerve

fibers misdirection (15). Thus, our results of a seemingly positive role of manual

mobilization on sciatic nerve regeneration add to those of the latter studies, and

reinforce the notion that passive exercise is an effective strategy to stimulate nerve

regeneration.

Despite the importance of nerve regeneration, a successful nerve treatment is much

dictated by the degree of reinnervation and of functional recovery. Here, we conducted

gait analysis to track functional recovery and assessed motion changes at the hip, knee

and ankle joints during voluntary level walking. Several motion changes occurred that

can be attributed to the sciatic nerve injury and that affected all recorded joints. In the

weeks following the sciatic nerve crush and before initiating either active or passive

exercise, relevant changes in hip and knee kinematics were noticed, particularly during

the stance phase. Higher hip joint velocity and overall increase in knee extension

characterized walking of sciatic-injured animals acutely post injury. This profound

reorganization of the role played by each individual joint reveals most likely a strategy

to overcome dysfunction at the level of ankle joint and to permit walking. Similar to

other studies (2; 32; 33), twelve weeks after sciatic injury ankle kinematics was recover

in all sciatic-injured groups, taking into account either the ankle angle and the ankle

angular velocity during both the stance and swing phases. Complete recovery of the

normal ankle kinematics during level walking in a walkway is a common observation

after a sciatic nerve crush, however some studies report few kinematic changes

appearing late during recovery suggesting that walking is altered after this type of injury

(32; 33). Also, it should be kept in mind that the morphology of regenerated crushed



sciatic nerves are clearly distinct from intact nerves, which suggest that complete

functional recovery is unlikely to occur on these animals. However, and in contrast to

ankle joint kinematics, hip and knee joint kinematics in the Crush group clearly deviated

to that of the Sham group. Such changes in the normal hip and knee kinematics during

walking were not noticed in both the EXa and EXp groups. We interpret these results

as a demonstration that the hindlimb motion pattern during level walking is not fully

reestablished after sciatic crush injury and those compensatory changes at the hip and

knee joint kinematics emerges that masks disability at the ankle joint. The mechanisms

guiding these changes cannot be discerned by our data, but studies have shown

altered pattern of electromyographical activity of the muscles crossing the ankle joint

after sciatic nerve transection that are best explained by changes at spinal neural

output (34). A possibility is that hip and knee motion pattern changes result from

altered proprioceptive feedback originating from muscles crossing the ankle joint and

other hindleg afferents (35; 36). A simpler explanation, and probably one that better

adjusts to our results, is that hip and knee kinematic changes reflect past

compensatory changes that are not corrected even if their purpose is not justified

anymore. In fact, the stimulus of active and passive exercise corrected hip and knee

motion changes during level walking in our exercised animals.

This study was not designed to investigate the mechanisms by which active or passive

exercise improves nerve regeneration and functional recovery. The mechanisms are

potentially many and might operate at several levels and at distinct organs, including

the spinal cord (19; 37), the regenerating nerve (Sabatier, Redmon et al. 2008) and the

reinnervated muscles (15; 38). The enhancing effect of active exercise on nerve

regeneration and reinnervation probably expresses the maintenance of an adequate

degree of stimulation onto the affected motor groups at the spinal cord level. This

stimulus might come either from the descending pathways that are active when

animals attempt to walk, even if due to denervation the movement itself does not

happen, or from afferent pathways originating from non-denervated hindlimb territories

but projecting onto the sciatic motoneuron pools (15; 39). It is known from cultured

retinal ganglion cells that neuronal outgrowth does not occur spontaneously and is

dependent on soluble neurotrophic factors and neuronal electrical activity (40). It is

likely that both active and passive exercise originate bursts of electrical activity at the

neuronal soma in the spinal cord that then travel along the severed axons and

stimulate axonal outgrowth.

The positive effect of passive exercise on nerve regeneration and functional recovery is

probably less anticipated, compared to active exercise. Previous studies used a

passive exercise stimulus that was designed to closely resemble natural active


crush injury 157


movements and behaviors. For instance, passive bicycling used by Udina et al. (16)

emulated the range of motion of hindlimb joints observed during walking in the rat.

Similarly, the manual strokes of the whisker pad implemented by Angelov et al. (15)

took into consideration the normal sweep exploratory motions of the vibrissae in these

animals. Our manual passive mobilization of the hindlimb was not planned to mimic

any specific task but otherwise to move the hip, knee and ankle joints through most of

their movement amplitude and to cause a mild stretch of the muscles crossing the

ankle joint. This protocol is in line with rehabilitation techniques aiming to maintain the

joint amplitude and prevent muscle shortening and contracture. However, the stimulus

caused by this rather unspecific passive motion proves also to have a positive effect on

the regenerating nerve morphology and in overall hindlimb motion pattern during

walking. Such apparent positive direct effect on nerve regeneration might result from

increased afferent input due to the movement stimulus (19; 37). Another possibility is

that our passive movement induced rhythmic mechanical load onto the regenerating

nerve inducing local cellular and molecular responses that enhance nerve regeneration

(41).

Lastly, some limitations of our study must be addressed. Our active walking exercise

might be envisaged as a mild-intensity endurance exercise, consisting of daily sessions

of 30 minutes continuous level walking at a 10 m/min speed. This exercising protocol

was initiated two weeks after sciatic nerve crush, when injured axons start

reinnervating the target muscles. However, there is evidence that the nerve

regeneration enhancing effect of exercise might be dose-dependent, being endurance

high-intensity, intermittent physical exercise capable of inducing higher effect than mild-

intensity (17). Also, some protocols use treadmill exercise of one-to two-hours duration

(13; 16; 31). It is then possible that our slight positive effect of active exercise on nerve

regeneration and functional recovery is the reflection of a relatively low exercise

stimulus. Also, our walking conditions used to test functional recovery might not

challenge the walking animals to a degree sufficient to uncover the disability associated

to faulty reinnervation or changes at spinal neural output (42).

In summary, active continuous walking exercise and passive mobilization of whole

hindlimb had a positive effect on the morphology of regenerating crushed sciatic nerves

and improved overall hindlimb kinematics during the rat gait. These results are in line

with growing evidence that not just active exercise but passive exercise as well are

suitable strategies to help nerve regeneration after injury and, most important, to aid in

function restoration. Moreover, the positive effect of passive exercise after sciatic nerve

injury does not require a strict task-specific movement and might alternatively result



from a general action of movement stimulation, including a direct mechanical effect on

the nerve and muscles. Clearly, further studies are needed for in vivo investigation of

the mechanism responsible for the effect of passive exercise.


crush injury 159


References

1. Raimondo S, Fornaro M, Tos P, Battiston B, Giacobini-Robecchi MG, Geuna S.

Perspectives in regeneration and tissue engineering of peripheral nerves.

[Internet]. Annals of anatomy = Anatomischer Anzeiger : official organ of the

Anatomische Gesellschaft. 2011 Mar 12;




;29(33):4409-19.

3. Malysz T, Ilha J, Nascimento PSD, Angelis KD, Schaan BD, Achaval M.

Beneficial effects of treadmill training in experimental diabetic nerve regeneration

[Internet]. Clinics. 2010 ;65(12):1329-1337.



Neurological research. 2004 Mar ;26(2):186-94.

5. Bodine-Fowler SC, Meyer RS, Moskovitz A, Abrams R, Botte MJ. Inaccurate

projection of rat soleus motoneurons: a comparison of nerve repair techniques.

[Internet]. Muscle & nerve. 1997 Jan ;20(1):29-37.[

6. Brushart TM. Preferential reinnervation of motor nerves by regenerating motor

axons. [Internet]. The Journal of neuroscience : the official journal of the Society

for Neuroscience. 1988 Mar ;8(3):1026-31.



in the rat sciatic nerve model. [Internet]. Experimental neurology. 2008 Jun

;211(2):339-50.

8. Bernstein JJ, Guth L. Nonselectivity in establishment of neuromuscular

connections following nerve regeneration in the rat. [Internet]. Experimental

neurology. 1961 Sep ;4262-75.

9. Valero-Cabré A, Navarro X. H reflex restitution and facilitation after different

types of peripheral nerve injury and repair. [Internet]. Brain research. 2001 Nov

23;919(2):302-12.

10. Meeteren NL van, Brakkee JH, Helders PJ, Gispen WH. The effect of exercise

training on functional recovery after sciatic nerve crush in the rat. [Internet].

Journal of the peripheral nervous system : JPNS. 1998 Jan ;3(4):277-82.



11. Marqueste T, Alliez J-R, Alluin O, Jammes Y, Decherchi P. Neuromuscular

rehabilitation by treadmill running or electrical stimulation after peripheral nerve

injury and repair. [Internet]. Journal of applied physiology (Bethesda, Md. :

1985). 2004 May ;96(5):1988-95.

12. Sabatier MJ, Redmon N, Schwartz G, English AW. Treadmill training promotes

axon regeneration in injured peripheral nerves [Internet]. Experimental

neurology. 2008 ;211(2):489–493.

13. Asensio-Pinilla E, Udina E, Jaramillo J, Navarro X. Electrical stimulation

combined with exercise increase axonal regeneration after peripheral nerve

injury. [Internet]. Experimental neurology. 2009 Sep ;219(1):258-65.




Nov ;517(2):245-55.

15. Angelov DN, Ceynowa M, Guntinas-Lichius O, Streppel M, Grosheva M,

Kiryakova SI, et al. Mechanical stimulation of paralyzed vibrissal muscles

following facial nerve injury in adult rat promotes full recovery of whisking.

[Internet]. Neurobiology of disease. 2007 Apr ;26(1):229-42.

16. Udina E, Puigdemasa A, Navarro X. Passive and active exercise improve

regeneration and muscle reinnervation after peripheral nerve injury in the rat.

[Internet]. Muscle & nerve. 2011 Feb ;1-10.




Nov ;517(2):245-55.

18. English AW. Enhancing axon regeneration in peripheral nerves also increases

functionally inappropriate reinnervation of targets. [Internet]. The Journal of

comparative neurology. 2005 Oct ;490(4):427-41.

19. Pavlov SP, Grosheva M, Streppel M, Guntinas-Lichius O, Irintchev A, Skouras

E, et al. Manually-stimulated recovery of motor function after facial nerve injury

requires intact sensory input. [Internet]. Experimental neurology. 2008 May

;211(1):292-300.






crush injury 161



2004 Nov ;21(11):1652-70.




methods. 2007 Jun ;163(1):92-104.




502.




review of neurobiology. 2009 Jan ;8781-103.[



Jan ;24(1):72-6.



myelinated nerve fibers in peripheral nerves. [Internet]. Annals of anatomy =

Anatomischer Anzeiger : official organ of the Anatomische Gesellschaft. 2000

Jan ;182(1):23-34.

27. Geuna S. The revolution of counting “tops”: two decades of the disector principle

in morphological research. [Internet]. Microscopy research and technique. 2005

Apr ;66(5):270-4.

28. Pakkenberg B, Gundersen HJ. Neocortical neuron number in humans: effect of

sex and age. [Internet]. The Journal of comparative neurology. 1997 Jul

28;384(2):312-20.

29. Nascimento PS do, Malysz T, Ilha J, Araujo RT, Hermel EES, Kalil-Gaspar PI, et

al. Treadmill training increases the size of A cells from the L5 dorsal root ganglia

in diabetic rats. [Internet]. Histology and histopathology. 2010 Jun ;25(6):719-32.

30. Malysz T, Ilha J, Nascimento PS do, Angelis KD, Schaan BD, Achaval M.

Beneficial effects of treadmill training in experimental diabetic nerve

regeneration. [Internet]. Clinics (São Paulo, Brazil). 2010 Jan ;65(12):1329-37.

31. Ilha J, Araujo RT, Malysz T, Hermel EES, Rigon P, Xavier LL, et al. Endurance

and resistance exercise training programs elicit specific effects on sciatic nerve



regeneration after experimental traumatic lesion in rats. [Internet].

Neurorehabilitation and neural repair. 2008 ;22(4):355-66.

32. Luis AL, Rodrigues JM, Lobato JV, Lopes MA, Amado S, Veloso AP, et al.

Evaluation of two biodegradable nerve guides for the reconstruction of the rat

sciatic nerve. [Internet]. Bio-medical materials and engineering. 2007 Jan

;17(1):39-52.




34. Sabatier MJ, To BN, Nicolini J, English AW. Effect of slope and sciatic nerve

injury on ankle muscle recruitment and hindlimb kinematics during walking in the

rat. [Internet]. The Journal of experimental biology. 2011 Mar 15;214(Pt 6):1007-

16.

35. Maas H, Prilutsky BI, Nichols TR, Gregor RJ. The effects of self-reinnervation of

cat medial and lateral gastrocnemius muscles on hindlimb kinematics in slope

walking. [Internet]. Experimental brain research. Experimentelle Hirnforschung.

Expérimentation cérébrale. 2007 Aug ;181(2):377-93.

36. Chang C-J. The effect of pulse-released nerve growth factor from genipin-

crosslinked gelatin in schwann cell-seeded polycaprolactone conduits on large-

gap peripheral nerve regeneration. [Internet]. Tissue engineering. Part A. 2009

Mar ;15(3):547-57.

37. Chen Y, Wang Y, Chen L, Sun C, English AW, Wolpaw JR, et al. H-reflex up-

conditioning encourages recovery of EMG activity and H-reflexes after sciatic

nerve transection and repair in rats. [Internet]. The Journal of neuroscience : the

official journal of the Society for Neuroscience. 2010 Dec 1;30(48):16128-36.

38. Cebasek V, Kubínová L, Janácek J, Ribaric S, Erzen I. Adaptation of muscle

fibre types and capillary network to acute denervation and shortlasting

reinnervation. [Internet]. Cell and tissue research. 2007 Nov ;330(2):279-89.

39. Udina E, Voda J, Gold BG, Navarro X. Comparative dose-dependence study of

FK506 on transected mouse sciatic nerve repaired by allograft or xenograft.

[Internet]. Journal of the peripheral nervous system : JPNS. 2003 Sep ;8(3):145-

54.

40. Goldberg JL, Espinosa JS, Xu Y, Davidson N, Kovacs GTA, Barres BA. Retinal

ganglion cells do not extend axons by default: promotion by neurotrophic

signaling and electrical activity. [Internet]. Neuron. 2002 Feb 28;33(5):689-702.

41. Heidemann SR, Buxbaum RE. Mechanical tension as a regulator of axonal

development. [Internet]. Neurotoxicology. 1994 Jan ;15(1):95-107.


crush injury 163


42. Sabatier MJ, To BN, Nicolini J, English AW. Effect of axon misdirection on

recovery of electromyographic activity and kinematics after peripheral nerve

injury. [Internet]. Cells, tissues, organs. 2011 Jan ;193(5):298-309.

Chapter 7 - General discussion and conclusions

166 Functional assessment after Peripheral Nerve Injury - Hindlimb Kinematic Model of the rat


1 General discussion

The main objective of this thesis was to investigate functional recovery of rats’ hindlimb

after therapeutic strategies for peripheral nerve repair in axonotmesis and neurotmesis

injuries. The results obtained in the experimental studies included in this thesis allowed

to verify functional and morphological differences between different therapeutic

strategies, when using tissue engineering and cellular therapies based on new

biomaterials. Chitosan (Chapter 3) and collagen (Chapter 4) associated to a cellular

system capable of neurotrophic factors production are the different therapeutic

strategies. Exercise-based neurorehabilitation was also studied in chapter 6.

The relevance of rat model in this thesis is related with the possibility to isolate several

levels of complexity to understand biological systems and develop cellular systems to

promote nervous tissue regeneration. It was possible to evaluate functional recovery in

vivo after therapeutic intervention for peripheral nerve repair during a follow-up period

with morphological evaluation of the nerve at the end of the experiment. Our

experimental model was the sciatic nerve, and we have verified that the possibility of

peripheral nerve repair is a promising reality. Peripheral nerve regeneration can be

studied in a number of different experimental models based on the use of nerves from

both forelimb and hindlimb (1-3). The experimental animal model of choice for many

researchers remains the rat sciatic nerve. It provides an inexpensive source of

mammalian nervous tissue of identical genetic stock that it is easy to work with and

well studied (4) and shows a similar capacity for regeneration in rats and sub-humans

primates (5). The rat sciatic nerve is a widely used model for the evaluation of motor as

well as sensory nerve function at the same time (3).

Our studies of sciatic nerve regeneration include a post-surgery follow-up period of 20

weeks after neurotmesis and 12 weeks after axonotmesis based on the assumption

that, by the end of this time, functional and morphological recovery are complete (6-9).

Although both morphological and functional data have been used to assess neural

regeneration after induced neurotmesis and axonotmesis injuries, the correlation

between these two types of assessment is usually poor (10-12). Classical and newly

developed methods of assessing nerve recovery, including histomorphometry,

retrograde transport of horseradish peroxidase and retrograde fluorescent labelling (13;

14) do not necessarily predict the reestablishment of motor and sensory functions (10;

15-17). Although such techniques are useful in studying the nerve regeneration

process, they generally fail in assessing functional recovery (10). In this sense,

Chapter 7- General discussion and conclusions 167

Sandra Crisitina Fernanades Amado

research on peripheral nerve injury needs to combine both functional and

morphological assessment. The use of biomechanical techniques and rat’s gait

kinematic evaluation is a progress in documenting functional recovery (17). Indeed, the

use of biomechanical parameters has given valuable insight into the effects of the

sciatic denervation/reinnervation, and thus represents an integration of the neural

control acting on the ankle and foot muscles (17-20).

After peripheral nerve injury, acute changes were particularly evident during push off

phase during stance, reflected in increased dorsiflexion at Toe Off (TO). With time,

there was a somewhat limited recovery towards normality during stance phase with

less pronounced dorsiflexion at TO, suggesting that extensor muscles regained the

ability to partially perform the push off action.

In sciatic-crushed (i.e. axonotmesis) animals, functional recovery is fast and by week 6

to 8 most of the kinematic parameters show values similar to those post-surgery.

However, some of the parameters remain altered even after a 12-weeks recovery

period which contrasts to other functional test that recover within 8 weeks following

sciatic crush injury (8).

After axonotmesis injury, hybrid chitosan membranes were used for peripheral nerve

regeneration. Chitosan has recently attracted particular attention because of its

biocompatibility, biodegradability, low toxicity, low cost, enhancement of wound-healing

and antibacterial effects. Its potential usefulness in nerve regeneration have been

demonstrated both in vitro and in vivo (7; 22). Chitosan is a partially deacetylated

polymer of acetyl glucosamine obtained after the alkaline deacetylation of chitin (23).

Chitosan matrices have been shown to have low mechanical strength under

physiological conditions and to be unable to maintain a predefined shape for

transplantation, which has limited their use as nerve guidance conduits in clinical

applications. The improvement of their mechanical properties can be achieved by

modifying chitosan with a silane agent. γ-glycidoxypropyltrimethoxysilane (GPTMS) is

one of the silane-coupling agents, which has epoxy and methoxysilane groups. The

epoxy group reacts with the amino groups of chitosan molecules, while the

methoxysilane groups are hydrolyzed and form silanol groups, and the silanol groups

are subjected to the construction of a siloxane network due to the condensation. Thus,

the mechanical strength of chitosan can be improved by the crosslinking between

chitosan and GPTMS and siloxane network. Chitosan and chitosan-based materials



have been proven to promote adhesion, survival, and neuritis outgrowth of neural cells

(7).

The normal pattern of ankle motion during stance characterized by initial dorsiflexion

followed by plantarflexion (push off action) was lost and dorsiflexion continues up to the

end of the stance phase caused by inability of the leg muscles to generate

plantarflexion moment (21). The ankle joint angle at Initial Contact changed

significantly after the sciatic crush. Such changes were transient and by week4 ankle

joint angle at Initial Contact was indistinguishable from baseline. Ankle joint angle at

Opposite Toe-off was significantly affected after the crush injury but by week 4 of

recovery joint position at this point of stance had recovered to normal values. At Heel

Rise ankle joint angle was significantly altered after the sciatic nerve injury until week

12. Ankle joint angle at Toe-Off was significantly altered after the sciatic nerve injury. At

week 12 ankle’s joint angle mean value at Toe-Off was similar to the one registered

before the sciatic nerve injury.

Group differences were observed, being the values at Initial Contact in the ChitosanII

group different from all the other four groups. Ankle joint angular position at Opposite

Toe in the ChitosanII group, throughout the study, were significantly different from

Crush and ChitosanIII groups. At Heel Rise, it was possible to verify differences

between ChitosanII and the Crush and ChitosanIIICell groups. Ankle’s joint angle at

Toe-Off was different between ChitosanII and ChitosanIIICell groups.

Ankle angular velocity at Opposite Toe was significantly affected after the crush injury

and returned to preoperative values at week 4. Instantaneous ankle’s joint velocity at

Heel Rise was significantly affected after axonotmesis with differences from normal

values until week 4 of recovery. Ankle’s joint velocity at Toe-Off was significantly

affected after the crush injury and did not recover completely within the 12-weeks

follow up time.

There were significant differences between the groups for ankle joint velocity at

Opposite Toe with significant differences between ChitosanII and ChitosanIICell

groups. Differences between ChitosanII and ChistosanIICell groups were registered at

Heel Rise. At Toe-Off, differences between ChitosanIICell, from one hand and Crush,

ChitosanIII and ChistosanIIICell groups, on the other were found. At week12, results of

the gait kinematic analysis of the experimental groups revealed that the group using



ChitosanIII group differs from Crush group in a smaller number of variables. This

indicates that at the end of the recovery period animals in ChitosanIII group presented

a degree of functional recovery similar to that registered in the Crush group.

Sciatic functional index (SFI) and Static Sciatic Index (SSI) results showed that SFI

values changed significantly after the crush injury. SFI values were different from

baseline until week 7 of recovery.

The effect of group was observed for SFI values indicating that values from

ChitosanIICell group were different from those of the ChitosanII and ChitosanIII groups.

Analysis on SSI data was similar to the reported SFI results. However, SSI results from

the ChitosanIIICell group were different from those of the Crush, ChitosanII and

ChitosanIICell groups.

The Extensor Postural Trust (EPT) values obtained during the healing period of 12

weeks were significantly affected by the crush injury and at week 12, EPT had still not

recovered to baseline values.

There were significant differences in EPT values between the experimental groups

analysis shows that the EPT values from the Crush group were significantly different

from those of the ChitosanIICell and ChitosanIIICell, suggesting a negative effect of the

N1E 115-differentiated cells on motor recovery after sciatic crush injury.

Withdrawal Reflex Latency (WRL) assessed nociceptive function and it was affected up

to week 8. Beyond this time point, WRL values, the experimental groups pooled

together, were similar to those preoperatively.

The effect of group was found for WRL data and analysis showed that differences were

significant between ChitosanII group from one side and ChitosanIICell, ChitosanIII and

ChitosanIIICell groups on the other side. These results are suggestive of a slower rate

of recovery in WRL with the use of type II chitosan.

Functional assessment partially supported the histomorphometric results since results

of the WRL, EPT, SFI and SSI tests suggest that chitosan type II and the N1E-115

cells have a reduced degree of functional recovery which is in line with results of

histomorphometric analysis. Histological appearance of control normal sciatic nerve

cross sections in comparison to cross sections of the regenerated nerve fibers in all

five experimental groups were organized in microfascicles and were smaller than

control nerve fibers. In the two experimental groups in which cell delivery was carried



out, the presence of unusual cell profiles, interpretable as the transplanted cells, was

detectable at the periphery of the nerves. This was seldom observed in inner parts of

the nerve suggesting that only few transplanted cells colonized the inner nerve

interstice. Electron microscopy confirms that a good regeneration pattern of both

unmyelinated and myelinated nerve fibers occurred in all experimental groups,

irrespectively of the enrichment with neural stem cells. The border zone of the nerve

still shows the presence of some chitosan debris integrated with the collagen fibers of

the epineurium. Moreover, in the two experimental groups in which cell delivery was

carried out, it was possible to find some of transplanted cells localized in the border

zone of the regenerated nerve. Results of the design-based morphoquantitative

analysis of regenerated myelinated nerve fibers permitted to quantitatively compare the

different experimental groups. As expected, fiber density was significantly higher in all

experimental groups in comparison to control sciatic nerves, while mean axon and fiber

diameter and myelin thickness were significantly lower. On what regards the number of

regenerated myelinated nerve fibers, was found to be significantly higher in comparison

to controls in four out of the five experimental groups only; in fact, ChitosanIII group

showed a number of fibers not significantly different in comparison to normal control

nerves. The peculiarity of the experimental group, characterized by type III chitosan

membrane enwrapment, was also confirmed by statistical analysis of inter-group

variability among experimental groups which demonstrated that ChitosanIII group

presented a significantly higher mean fiber and axon diameter and myelin thickness in

comparison with all other experimental crush groups.

We also studied the effects on nerve regeneration when using collagen membranes

after neurotmesis without gap (end-to-end microsurgical technique) associated with in

vitro differentiated cellular system (chapter 3). After 20 weeks animals were sacrificed

and the repaired sciatic nerves were processed for histological and stereological

analysis. SFI was not possible to perform because of automutilation of the fingers after

sciatic nerve injury.

In the weeks following sciatic nerve transection, ankle joint motion became severely

abnormal, particularly throughout the second half of stance corresponding to the push-

off sub-phase. In clear contrast to the normal pattern of ankle movement, at week-2

post-injury animals were unable to extend this joint and dorsiflexion continued

increasing during the entire stance. This result could be associated to the paralysis of

plantarflexor muscles. The pattern of the ankle joint motion seemed to have improved

only slightly during recovery. For OT velocity and HR angle no differences were found



before and after sciatic nerve transection, whereas for OT angle differences from pre-

injury values were significant only at weeks 2 and 16 of recovery. The angle at IC

showed a unique pattern of changes, being unaffected at week-2 post-injury and

altered from normal in the following weeks of recovery. Probably the most consistent

results are those of HR velocity, TO angle and TO velocity. These parameters were

affected immediately after the nerve injury and remained abnormal along the entire 20-

weeks recovery period.

Generally, no differences in the kinematic parameters were found between the groups.

Exceptions were IC velocity in the End-to-EndMembCell group, which was different

from the other two groups, and OT angle in the End-to-EndMemb group that was also

different from the other two groups.

Reflex activity also revealed relevant information concerning functional recovery. The

EPT response steadily improved during recovery but at week-20 the EPT values of the

injured side were still significantly lower compared to values at week-0.

A significant main effect for treatment was found, with better recovery of the EPT

response in the End-to-EndMembCell group when compared to the other two

experimental groups: at week-20, motor deficit decreased to 27% in the End-to-

EndMembCell and to 34% and 42% in the End-to-End and End-to-EndMemb groups,

respectively.

During the following weeks after injury there was recovery in paw nociception, which

was more clearly seen between weeks 6 and 8 post-surgery. At week-6, half of the

animals still had no withdrawal response to the noxious thermal stimulus in the

operated side, which is in contrast with week-8, when all animals presented a

consistent, although delayed, response. Despite such improvement in WRL response,

persistence of sensory deficit in all groups by the end of the 20-weeks recovery time

was reported. No difference between the groups was observed in the level of WRL

impairment after the sciatic nerve neurotmesis.

After neurotmesis, regeneration of axons was organized in many smaller fascicles in

comparison to controls. No significant difference regarding any of the morphological

parameters investigated in the regenerated axons from the three experimental groups.

On the other hand, comparison between regenerated and control nerves showed, as

expected, the presence of a significantly higher density and total number of myelinated



axons in experimental groups accompanied by a significantly lower fiber diameter.

Results showed that enwrapment of the rapair site with a collagen membrane, with or

without neural cell enrichment, did not lead to any significant improvement in most of

functional and stereological predictors of nerve regeneration. The exception was EPT

which recovered significantly better after neural cell enriched membrane employment.

Results of this study revealed that this particular type of nerve tissue engineering

approach has very limited effects on nerve regeneration after sciatic end-to-end nerve

reconstruction in the rat.

We have concluded that the sciatic nerve regeneration after neurotmesis and end-to-

end suture can be improved by in vitro differentiated N1E-115 neural cells seeded on a

type III equine collagen membrane, enwrapped around the injury site. The N1E-115

cell line has been established from a mouse neuroblastoma (24) and have already

been used, with conflicting results due to its neoplasic origin, as a cellular system to

locally produce and deliver neurotrophic factors (25). In vitro, the N1E-115 cells

undergo neuronal differentiation in response to dimethylsulfoxide (DMSO), adenosine

3’, 5’- cyclic monophosphate (cAMP), or serum withdrawal (24). Upon induction of

differentiation, proliferation of N1E-115 cells ceases, extensive neurite outgrowth is

observed and the membranes become highly excitable (8; 9; 26-30). The interval

period of 48 hours of differentiation was previously determined by measurement of the

intracellular calcium concentration (Ca2+i). Notice that at this time, the N1E-115 cells

present, already, the morphological characteristics of neuronal cells, nevertheless cell

death due to increased Ca2+i is not yet occurring as described elsewhere (8; 9; 26;

29-36).

Neurotrophic factors play an important role in nerve regeneration after injury or disease

and it is conceivable that if neurotrophic factors are applied in the close vicinity of the

injured nerve their healing potency is optimized. In spite of these assumptions and

contrary to our initial hypothesis, the N1E-115 cells did not facilitate either nerve

regeneration or functional recovery and, as far as morphometrical parameters are

concerned, results showed that the presence of this cellular system reduced the

number and size of the regenerated fibers. These results suggest that this type of

nerve guides can partially impair nerve regeneration, at least from a morphological

point of view (8; 9; 26; 29-36). The impaired axonal regeneration seems to be the result

of N1E-115 cells surrounding and invading the regenerating nerve, since numerous of

these cells where seen colonizing the nerve and might have deprived regenerated

nerve fibers blood supply (8; 9; 26-30). The use of N1E-115 cells did not promote nerve



healing and their use might even derange the nerve regenerating process. Whereas

the effects on nerve regeneration were negative, an interesting result of this study was

the demonstration that the cell delivery system that we have used was effecting in

enabling long-term colonization of the regenerated nerve by transplanted neural cells.

Whether the negative effects of using N1E-115 cells as a cellular aid to peripheral

neural tissue regeneration extends to other types of cells is not known at present and

further studies are warranted to assess the role of other cellular systems, e.g.

mesenchymal stem cells, as a foreseeable therapeutic strategy in peripheral nerve

regeneration. Our experimental results with this cellular system are also important in

the perspective of stem cell transplantation employment for improving posttraumatic

nerve regeneration with patients (8; 9; 26; 29-36). Undoubtedly, great enthusiasm has

raised among researchers and in the general public about cell-based therapies in

regenerative medicine (8; 9; 26; 29-36). There is a widespread opinion that this type of

therapy is also very safe in comparison to other pharmacological or surgical therapeutic

approaches. By contrast, recent studies showed that cell-based therapy might be

ineffective for improving nerve regeneration (8; 9; 26; 29-36) or even have negative

effects by hindering the nerve regeneration process after tubulisation repair. Whereas

the choice of the cell type to be used for transplantation is certainly very important for

the therapeutic success, our present results suggest that the paradigm that donor

tissues guide transplanted stem cells to differentiate in the direction that is useful for

the regeneration process it is not always true and the possibility that transplanted stem

cells choose another differentiation line potentially in contrast with the regenerative

process should be always taken in consideration.

From a functional point of view, it was verified that independently from the type of

sciatic nerve injury, ankle motion is severely affected in the weeks immediately after

the injury. After neurotmesis (chapter 4), automutilation, chronic foot deformities,

inversion or eversion deformations were the main methodological source of limitation

for the functional assessment with the use of others widely used behavioural tests, i.e.

Sciatic Functional Index (SFI) and Static Sciatic Index (SSI) (12; 37). Despite this

limitation SFI is a widely used parameter because of its reliability (38), but recent

studies revealed that SFI method is only reliable from the 3rd week on after a severe

lesion (39). Although SFI is a quantitative method, it only considers stance phase and

several aspects have been reported as methodological limitations since some of the

measurements (i.e. Print Length factor) are dependent on the walking velocity and

pressure exerted by the foot on the floor influenced for instance by the weight of the



animal. Performing kinematic analysis for assessment of gait as dynamic function, as it

describes the trajectory of a joint during the movement, was a good alternative.

Although kinematics are only the description of movement and only takes into

consideration joint angles, it represent a good approach to understand joint

coordination and changes in kinematics during walking could be indicative of the role of

muscles in regulating locomotor activity. Using 2D video analysis and dedicated

software for motion analysis, our group reported measures of both angle and angular

velocity of the ankle joint during the stance phase (8; 26). Angular velocity data were

calculated in an attempt to raise the precision of joint motion analysis and to increase

its power in detecting subtle differences in functional recovery when testing alternative

treatments after sciatic crush (26). We reasoned that functional deficits during walking

in rat nerve models might be masked by the high redundancy and adaptability of the

motor apparatus in response to sensorimotor alterations (12; 37).

Almost all the kinematics parameters analysed regarding ankle motion during the

whole phase of the rat walk reveals a persistent abnormal pattern. Muscles innervated

by sciatic nerve rami includes both dorsiflexors and plantarflexors and although in

previous studies we focused our kinematic analysis only in the stance phase, including

analysis of all joints also during the swing phase provided additional information.

However, a key question regarding joint motion analysis is which parameters present

enough sensitivity to evaluate function after different types of sciatic injury. We have

evaluated the sensitivity and the specificity of ankle motion analysis to detect functional

deficits in the rat sciatic model (Chapter 5). By using a set of kinematic variables

describing angle and velocity during walking, a statistical model was generated that

robustly separates animals shortly after sciatic crush from sham-operated controls and

animals after 12 weeks of recovery. Moreover, the model displayed a moderate ability

to separate between the two latter groups of animals. Therefore, 2D kinematic analysis

of joint motion is sensitive to detect motor deficit caused by sciatic nerve injury even

when functional deficits are minimized by long-term recovery.

All kinematic measures were significantly altered in the time shortly after the sciatic

crush, when compared to sham-operated and recovered animals, with the exception of

angular velocity at mid-swing, which apparently is not affected by the sciatic nerve

crush and therefore seems not to be of great utility to assess functional status in the rat

sciatic model. Moreover, the introduction of interjoint coordination was also important in

order to have insights about the behaviour of all hindlimb joints, considering the

denervated chain of muscles after injury. Sciatic nerve gives off branches to the hip



extensors and leg flexors and continues its courses beneath the gluteus medius

muscle into the thigh, in the mid-thigh the nerve trifurcates into the peroneal, tibial and

sural branches; the peroneal nerve subsequently innervates the tibialis anterior and

extensor digitorum longus, while the tibial nerve supplies the plantar flexor, toe flexors,

and the tibialis posterior. Thus, it is expected that after sciatic nerve injury hip, knee

and ankle motion reflect functional deficits. Although kinematics is only the description

of movement and only takes into consideration joint angles, it represents a good

approach to understand joint coordination. Recent studies revealed that different joint

positions were combined to achieve functional movement during cat walking after

peripheral nerve injury (12; 37). The whole limb position and orientation were

considered a preferable parameter than joint position (38). Therefore, inter segmental

transfer of energy might also be a mechanisms of economical movement production

after muscle denervation.

Our data suggest that joint kinematics has the ability to highlight minor walking

changes after long-term recovery from sciatic crush injury. However, the kinematics

changes during walking twelve weeks after sciatic nerve crush were noticed in the hip

and knee joints, while ankle joint kinematics had recovered its normal pattern. In line

with joint kinematics data, discriminant analysis clearly demonstrated that walking

measures, either spatiotemporal features or hindlimb joint kinematics, accurately

recognize sciatic-denervated animals and distinguishes them from sciatic-reinnervated

animals and sham controls. Spatiotemporal parameters like gait velocity, stride length,

stance phase duration and swing phase duration are relevant for the study of

pathological pattern of gait after peripheral nerve injury in experimental rat model (39).

Stance phase duration measured the time hind foot was in contact with the floor. After

sciatic nerve injury, the injured limb has a smaller stance phase duration. Step length,

is accurately measured using the location of the middle metatarsal head to metatarsal

head in successive steps (8; 26) and walking speed is considered a parameter that

influences stance phase and step length after sciatic crush injury (26).

The relevance of spatiotemporal parameters is related with mechanical aspects of

locomotion and its functional relevance for performance. Considering that the basic

functions of locomotion are propulsion, stance stability, shock absorption and energy

conservation (40; 41), gravitational forces are important in walking and gait

disturbances reduces the ability to use gravity and inertia in ambulation and therefore

must resort to actual muscle work, which means that they have to spend more energy

to cover certain distance.



Finally, it was possible to detect subtle differences between two types of therapeutic

exercise widely used in neurorehabilitation clinical practice. The two exercise modes

consisted on treadmill exercise walking and passive mobilization of the affected

hindlimb and their effect was evaluated in contrast to a non-exercise group of sciatic

nerve crushed animals and sham operated controls. Passive exercise consisted on

manual mobilization of the entire right hindlimb. The manual mobilization included

movement of the hip, knee and ankle joints in concert, and alternating flexion and

extension of the joints. The results demonstrate marginal effects of both active and

passive exercise on functional recovery, assessed by hindlimb joint kinematics during

level walking. Although the impact of exercise on functional recovery was relatively

modest it was nonetheless accompanied by improved nerve morphology, particularly

by preventing excessive increase in total number of myelinated fibers in regenerated

sciatic nerves.

Kinematic analysis distinguished recovery of movement pattern between groups.

Twelve weeks after injury, ankle joint was similar between groups. In contrast to ankle

joint kinematics, hip and knee joint kinematics in the Crush group clearly deviated to

that of the Sham group. Such changes in the normal hip and knee kinematics during

walking were not noticed in both the active and passive exercise groups. We interpret

these results as a demonstration that the hindlimb motion pattern during level walking

is not fully reestablished after sciatic crush injury and that compensatory change at the

hip and knee joint kinematics masking disability at the ankle joint. The mechanisms

underlying these changes cannot be discerned by our data. The stimulus caused by

this rather unspecific passive motion proves also to have a positive effect on the

regenerating nerve morphology and in overall hindlimb motion pattern during walking.

Such apparent positive direct effect on nerve regeneration might result from increased

afferent input due to the movement stimulus (38). Another possibility is that our passive

movement induced rhythmic mechanical load onto the regenerating nerve inducing

local cellular and molecular responses that enhance nerve regeneration (39)

Comparison between regenerated and non-injured nerves showed morphological

differences. Signs of better regeneration as a result of passive hindlimb mobilization

were observed. As for the active exercise group, mean total number of myelinated

fibers in the passive exercise group was indistinct from that of the Sham group.

Interpretation of these results is difficult, and thus it deserves further research, the



discrepancy between different authors who revealed an increase in the number of

myelinated axons and our results, which evidenced the opposite. Thus, results from

exercise-based neurorehabilitation are promising. In conclusion, passive and active

joint movement causes joint kinematics and nerve morphological changes. Results of

this study suggest that mechanical load plays a role on functional recovery after

peripheral nerve repair. However, further studies are needed to understand the

functional meaning of results.



2 Future perspectives

2.1 Nerve regeneration

Enwrapment of the site of a nerve crush lesion with a chitosan type III membrane

significantly improves nerve regeneration in the rat, whereas enrichment of chitosan

membranes with N1E-115 neural cells does have positive effects. Whereas

transposition of animal studies to patients should always be dealt with caution, the

results obtained in this experimental study opened interesting perspectives for the

clinical use of hybrid chitosan membranes in peripheral nerve reconstruction.

More complex devices will be needed, such as multilayered tube guides where growth

factors are entrapped in polymer layers with varying physicochemical properties or

tissue engineered tube guides containing viable stem cells (8; 26). The combination of

two or more growth factors will likely exert a synergistic effect on nerve regeneration,

especially when the growth factors belong to different families and act via different

mechanisms. Combinations of growth factors can be expected to enhance further

nerve regeneration, particularly when each of them is delivered at individually tailored

kinetics (26). The determination and control of suitable delivery kinetics for each of

several growth factors will constitute a major hurdle both technically and biologically

with the biological hurdle lying in the compliance with the naturally occurring cross talk

between growth factors and cells. A solution to this problem may be the use of

mesenchymal autologous or heterologous stem cells because they can synthesize

several growth factors and differentiate into Schwann cells which are critical for very

long gaps (40; 41).

Tissue engineering advances promotes the development of other biomaterials to

enhance peripheral nerve regeneration. The choice of the cellular system to be applyed

is crucial for the therapeutic success. Using another cell line other than N1E-115 could

have led to better results, concerning for instance, the emerging knowledge about

autologous or heterologous mesenchymal stem cells from extra-fetal sources (38).

Moreover, the construction of more appropriate tube-guides with integrated growth

factor delivery systems and/or cellular components could improve the effectiveness of

nerve tissue engineering. In fact, single-molded tube guides may not give sufficient

control over both the mechanical properties and the delivery of bioactive agents.



2.2 Methodological improvement of functional assay

Kinematic analysis has improved the sensitivity of functional recovery assessment after

peripheral nerve injury in experimental animals and its importance seems will increase

in the future as a result of advanced technology and improved biomechanical models.

From a biomechanical perspective, joint rotational velocity has a more direct

relationship with the forces actuating onto the hindlimb segments and therefore velocity

measures may be better indicators of dysfunction caused by nerve injury. Moreover,

ankle position data is sometimes difficult to interpret, for example in those cases where

ankle joint angle remains unaffected in the weeks immediately after sciatic nerve crush

(8; 26). This shows that measures of ankle joint angle taken at snapshots during

walking may lack sensitivity to assess functional impairment. In the near future, other

statistical methods should be considered. Individual joint kinematics either in control or

nerve-injured animals is characterized by high variability, with notable differences

between different animals and even from step to step (42). Such high level of

variability, which seems to be an intrinsic property of normal quadruped walking,

seriously affects the precision of joint kinematic measures of functional recovery after

nerve injury. Reducing this variability is a challenge for efficient use of walking analysis

to assess functional recovery. Attempts to overcome this limitation include constraining

walking velocity by using treadmill walking instead of self-paced locomotion (28; 43-

48). This, of course, is likely to reduce step-by-step variability in joint kinematics but

has the disadvantage of requiring expensive equipment and limits the possibility of

combining kinematic analysis with other data, such as ground reaction forces. Other

possibilities look at a global, limb-level movement analyses as an alternative to

individual joints kinematics (28; 43-48). Also, systematic changes in the biomechanical

and movement control constraints of the locomotor task, such as using up- and down-

slope walking might also increase the accuracy of walking analysis within the context of

peripheral nerve research (28; 43-48).

Study the influence of gait velocity on functional parameters. Yu et al (2001) (37)

showed that gait analysis method is highly sensitive and studied seven spatial and

kinematic parameters. It was also pointed out that when the spontaneous walking

speed exceeded 55 cm/s was approached the limit of walking. When the walking speed

was less than 25cm/s the rats looked nervous and overcautious. Moreover, the stance

phase and step length, accurately measured by middle metatarsal head to metatarsal

head distance, was affected by walking speed.



Bilateral hindlimb model should be constructed. Yu et al. (2001) defined the right:left

stance/swing ratio (SS); right:left step length ratio (SLR); angle of the ankle joint at

terminal stance (ATS); angle of the ankle joint at midswing (AMS); tail height; tail

deviation, and midline deviation (MID). Concerning to tail and midline deviation, they

verified that the tail of normal rats deviated towards the side where the hindlimb was at

terminal stance, and during walking the body leans towards the contralateral side when

the hindlimb is at midswing. During a gait cycle after sciatic nerve injury produces ankle

dorsiflexion deformity presumably due to the dominance of tibial nerve innervation in

the lower extremities. Based on discriminant analysis these authors concluded that

right:left stance/swing ratio, right:left step length ratio and ankle joint at terminal stance

contributed significantly to the equation developed to obtain sciatic injury score.

Methodological errors related with skin movement artefacts should be studied in

kinematic analysis with optimization methods. In fact, the main sources of error

affecting in vivo estimation of the pose of a skeletal bone with a non invasive skin

markers technique are related with skin slippage, namely soft tissue artefacts (37),

anatomical landmarks misallocation and instrumental noise (40; 41), and efforts have

been made to overcome this limitation. While the estimation of instrumental errors is

not problematic, the in vivo quantification of soft tissue artefacts and their effect on the

determination of the bone position and orientation is still an open issue. Rats have

excessive skin mobility that can generate large soft tissue artefacts therefore

increasing the kinematic and kinetics data error during motion analysis. Thus, it was

verified that using tattooing the anatomical landmarks misallocation and consequent

inter-test variability was controlled (50). Soft tissue artefacts were verified to be

problematic at the knee joint (51). The position of the knee joint is loosed by skin

coverage; (8; 26) proposed that it could be calculated by using hip and ankle joint

position and external individual measurements of femur and tibia length. (26), used

high speed X-ray video and reported that the largest error between skin and bone

derived angles occurred immediately before the paw contact, and at toe-off the error

was the smallest. It was also reported that skin-derived kinematics overestimated

angular values and the knee position was the joint most affected. The choice of skin

marker placement is a critical aspect for discussion about motion analysis, and

comparisons between studies should consider this aspect when interpreting for

instance joint angle values. Sagittal values derived from skin markers are only

approximation of the actual movements of the bones that comprises a joint of interest.



Another problem in kinematic studies related with protocol is the accurate

determination of cycle landmarks such as Initial contact and toe-off. Ground reaction

forces will improve the accuracy of determination of stance phase events. The

introduction of optoelectronic motion capture system (Chapter 4) improved accuracy of

the kinematic model, allowing instantaneous position estimation and reducing the error

introduced by digitalization process, improving data processing. Position estimation is

performed automatically and when a marker is missing it is estimated with

mathematical algorithm considering the coordinates of the following frames. Data using

this system has surprisingly showed that unlike our previous stance phase based

results showing significant improvement in ankle joint motion after sciatic injury,

analysis of swing phase shows that the abnormal ankle motion pattern remains without

significant improvement over the entire recovery period. Using optoelectronic

technology it will be possible to introduce anthropometric characteristics. In what

concerns biomechanical model, the relative position of anatomical segments is

important to perform biomechanical modelling. Reconstruction and analysis of in vivo

skeletal system kinematics using optoelectronic stereophotogrammetric data will allows

the joint kinematic description providing segment morphology (40; 41). Anatomical 3D

analyses (42) will be possible to quantify observable functional deficits in others planes

of movement further than sagittal, like rotation of the injured paw, curvation and

inversion of the feet (37; 49).

2.3 Functional meaning

From animal models we know that neural control of locomotion operates automatically

at a basic level, without conscious perception. There is a spinal network, which

generates a detailed muscle-specific spatio-temporal pattern similar to what is

observed during walking. However, this automatic level does not operates in isolation,

the final muscle activation pattern is also influenced by sensory feedback from joint,

muscle, and skin receptors (37). During ongoing movement fundamental mechanical

variables of force, length, and velocity are monitored within muscles by Golgi tendon

organs and by muscle spindle receptors, respectively. Thus, kinematics description of

motor behaviour provides a foundation for integrating morphology and physiology with

mechanics and function (28; 43-48). Moreover, force generated by a muscle depends

on activation by nerve impulses as well as the length and velocity of the muscle. The

functional meaning of recovery after peripheral nerve injuries implies the study of motor

control since peripheral nerves enclose both sensory and motor information, which are

important to produce controlled movement. From our results, we demonstrated that



joint angular displacement and joint velocity during locomotion movement might

quantify functional recovery after peripheral nerve injury. Moreover, nociception was

the first variable achieving signs of recovery. How could these quantities be related

with neuromuscular function and have functional meaning and clinical implications?

When we analyze joint kinematics data, it should be taken into consideration that these

results are also related with neural and muscle function thus being relevant to perform

a comprehensive description of movement of individual muscle action within the joint or

the joints it crosses. Dynamics of muscle function is usually examined by indirect

assessment of muscle length change based on kinematics (51) and muscle moment

arms, or estimates of fascicle length change that must take into account in-series

elastic stretch of a tendon and a simplification of the effects of muscle architecture.

Knowledge of the more complex, but a realistic pattern of 3D movement is important for

understanding its neuromuscular control, as well as the principles of musculoskeletal

design (52) that are crucial to the production of movement. During dynamic conditions,

combining kinematic data with kinetics and electromyography would be desirable to

make a comprehensive understanding of movement patterns and its changes as well

as the study of inefficient movement after injury. The study of the relationship between

joint angular velocity and muscle moments would contribute to understand muscle

contraction and work production. Results from the last study awake us for the

relevance of mechanical load, fatigue and the influence of joint rotational velocity,

which should be explored to understand performance and make available results

translation for motor rehabilitation.

Understanding functional results after peripheral nerve injury is complex, especially

when both sensory and motor pathways are impaired. Moreover, some limitation of

WRL and EPT should be considered and the interpretation of functional results should

be made carefully. Motor reflex deficit was assessed by extensor postural test (EPT)

that is characterized by the exertion of force on the ground, in a hand-controlled

loading condition, quantified with a scale. The reduction in this force, representing

reduced extensor muscle tone, was considered a deficit of motor function. Moreover,

the muscles that has been considered for interpretation of the results was the ankle

extensors i.e. gastrocnemius and soleus muscles. Also for sensory function, some

authors considered that the influence of mecanoreceptors should be controlled for the

interpretation of the results controlling the pressure exerted on the thermal platform.

Although WRL reflex was originally named flexion-withdrawal reflex (42), it has since

been shown to involve other muscles besides flexors (52). The WRL can vary because



it depends on which afferents are stimulated and signals transmitted over polysynaptic

pathways. This means that the input signal can be modified along its path. The

recovery of sensory function is undertaken by several afferent fibers and may not be

explained by the outgrowth of regenerating axons from the sectioned proximal sciatic

nerve stumps but also by collateral sprouts from intact fibers in the skin surrounding the

denervated zone. In fact, the recovery of sensory nerve function might be started with

the expansion of the territory of intact neighbouring fibers (28; 43-48). Regenerated

sciatic nerve began to regain function between 3 (28; 43-48) to 7 (52) weeks after

injury. The borderline between a sensory-motor response involving both sensory

ascending and descending motor pathways and a pain-based withdrawal reflex-

response is not always possible. It cannot always be sure that all responses are based

on proper temperature sensing. Reflex activity should be integrated in a dynamic

model, assessed during movement, since it might comprise an important aspect on

motor control with functional meaning for the position of the limb and movement.

Mathematical modelling is a possible solution to simplify and understand the potential

environmental effect with similar neural control and exclude neural feedback. Efforts

will be made to improve the biomechanical modelling with musculoskeletal model and

methodologies to understand the reflex mechanisms involved in the sensory and motor

recovery of function.

In summary, gait analysis is a promising method to assess functional recovery after

hindlimb nerve injury. However, in order to provide accurate measures of functional

recovery, gait analysis after hindlimb peripheral nerve injury should evolve from a

simple ankle kinematics analysis to a full 3D biomechanical description of complete

hindlimb motion, meaning analysis of hip, knee and ankle joints. Further refinements of

gait analysis in the field of peripheral nerve research using the rat model should include

the combined use of joint kinematics, ground reaction forces and electromyography

data.

Innervation regulates muscle mass and muscle phenotype (53) and peripheral nerve

injury in the rat is a widely used model to investigate nerve regeneration and can also

be employed as a model of muscle inactivity and muscle atrophy (54). Previous work of

our research team has been devoted to enhance nerve regeneration after injury,

including the use of biomaterials and cellular systems (26). By combining motor and

sensory function tests, biomechanical analysis of the rat gait and nerve morphology

assessment by unbiased stereological methods, we could demonstrate that animals



almost fully recover from sciatic nerve crush in 12 weeks (8). However, nerve

morphology remains significantly different in control and injured animals as well as

ankle joint motion during walk (8). Muscle function was not assessed in our previous

work. Aside incomplete nerve regeneration, changes in the muscles may contribute to

functional deficit after nerve injury (54-56).

Denervation induces muscle atrophy and, 25 months post denervation, muscle fibres

crossectional area of the EDL muscle diminish to only 2.5% of control animals although

their fascicular organization is maintained (57). The effect of denervation on muscle

atrophy is both activity-dependent and activity-independent since the degree of

hindlimb muscle atrophy after spinal isolation (activity-independent nerve influence) is

less when compared to the atrophy caused by removal of all nerve influences by

transecting the sciatic nerve (53). Two basic mechanisms are responsible for

denervation-induced muscle atrophy. First, there is augmented activity of the ubiquitin-

proteasome pathway and proteolysis (58). Second, there is cell death and myonuclei

apoptosis conjugated with decreased capacity of satellite cell-dependent reparative

myogenesis (54; 57). Together with atrophy, denervated/reinnervated muscles undergo

phenotypical changes and conversion between muscle fibre types (55).

The relative increase in type I or type II muscle fibres following denervation seems to

depend on the type of muscle fibres predominant in the muscle, with type II muscle

fibres (fast fibres) increasing in proportion in soleus (slow muscle) and type I muscle

fibre number increasing in gastrocnemius and tibialis anterior muscles (54). Likewise,

the degree of muscle fibre atrophy in short-term denervation (4 weeks) has been

noticed to be greater in the muscle fibre type that is more abundant in the affected

muscles (59). However, it must be stressed that for many purposes there is a lack of

detailed morphological description of denervated/reinnervated muscles since in most

cases the morphoquantitative methods employed lack accuracy and completeness.

Muscle architectural changes, and also likely remodeling of the enveloping connective

tissue (60) might contribute to altered function in denervated/reinnervated muscles.

Ideally, muscle contractile properties are evaluated in minimally dissected muscle-

nerve preparations in which muscle contractions are elicited by electrical stimulation of

the nerve (61). Using such approach, fundamental properties of hindlimb muscles of

the rat, like isometric peak force, range of active force, slack length, and passive

length-force curve can be obtained. All these parameters are potentially affected by

changes in muscle morphology secondary to denervation/reinnervation (60). Unlike the

classical view, skeletal muscles are not isolated units and their mechanical properties

are dependent on the context, meaning the mechanical load in surrounding muscles



and connective tissue (62). The transmission of force out from the muscle fibres to the

skeleton follows both myotendinous and myofascial links (63). The latter involves load

being transmitted to synergistic and antagonistic muscles (60) and occurs through the

muscular and non-muscular connective tissues (e.g. neurovascular tracts, i.e. the

collagen fiber reinforced tissues surrounding nerves and blood vessels). Surgical

reconstruction of the peripheral nerves affects the mechanical properties of the

neurovascular tract supplying leg muscles in unknown ways as well as the architecture

of denervated muscles. In addition, altered properties of supportive connective tissue

may affect myofascial force transmission and overall function (60). In situ testing of

muscle function, keeping the connective tissue that envelops the muscle or muscle

groups undamaged, is a unique method to evaluate possible changes in myofascial

force transmission (62), and by measuring the force produced at the proximal and

distal ends of the EDL an indication of myofascial force transmission is obtained (64).

In recent years, great emphasis has been placed on kinematic analysis of the rat

locomotion [(19). Although gait analysis is an elegant and meaningful way to assess

functionality, it is limited in evaluating physiological and mechanical properties of the

affected muscles. The potential for biomechanical gait analysis to assess hindlimb

muscle function requires precise motion capture system combined with ground reaction

force data (65) and a geometrical model of the rat hindlimb musculoskeletal system

(66).

The development of biological and material therapeutic strategies to enhance

functional recovery is the major interest of several research areas such as Biology,

Engineering, and Medicine. The assessment of functional recovery is a complex

question that has many different facts to be addressed in detail. A common feature of

these problems is that they address relationships between morphological, physiological

and behavioural aspects of information processing in nervous system (“structure-

function” relationships). Movement Sciences should be considered for the

interdisciplinary approaches across established neuroscience disciplines. Moreover,

functional recovery after peripheral nerve injury assessed with dynamical model of

locomotion might represent an integrative approach to have deep knowledge about

neural activity, mechanical and structural relationship.



REFERENCES

1. Yang C-C, Shih Y-H, Ko M-H, Hsu S-Y, Cheng H, Fu Y-S. Transplantation of

human umbilical mesenchymal stem cells from Whartonʼs jelly after complete

transection of the rat spinal cord. [Internet]. PloS one. 2008 Jan ;3(10):e3336.

2. Höke A. Mechanisms of Disease: what factors limit the success of peripheral

nerve regeneration in humans? [Internet]. Nature clinical practice. Neurology.

2006 Aug ;2(8):448-54.

3. Jensen JN, Tung THH, Mackinnon SE, Brenner MJ, Hunter DA. Use of anti-

CD40 ligand monoclonal antibody as antirejection therapy in a murine peripheral

nerve allograft model. [Internet]. Microsurgery. 2004 Jan ;24(4):309-15.

4. Geuna S, Raimondo S, Ronchi G, Di Scipio F, Tos P, Czaja K, et al.

International Review of Neurobiology Volume 87 [Internet]. Elsevier; 2009.


after sciatic nerve blockade. [Internet]. Anesthesiology. 1997 Apr ;86(4):957-65.[

6. Amado S, Rodrigues JM, Luís AL, Armada-da-Silva P a S, Vieira M, Gartner A,

et al. Effects of collagen membranes enriched with in vitro-differentiated N1E-

115 cells on rat sciatic nerve regeneration after end-to-end repair. [Internet].

Journal of neuroengineering and rehabilitation. 2010 Jan ;77.

7. Simões MJ, Amado S, Gärtner A, Armada-Da-Silva P a S, Raimondo S, Vieira

M, et al. Use of chitosan scaffolds for repairing rat sciatic nerve defects.

[Internet]. Italian journal of anatomy and embryology = Archivio italiano di

anatomia ed embriologia. 2010 Jan ;115(3):190-210.




methods. 2007 Jun ;163(1):92-104.




;28(6):458-70.

10. Shen N, Zhu J. Application of sciatic functional index in nerve functional

assessment. [Internet]. Microsurgery. 1995 Jan ;16(8):552-5.

11. Kanaya F, Firrell JC, Breidenbach WC. Sciatic function index, nerve conduction

tests, muscle contraction, and axon morphometry as indicators of regeneration.



[Internet]. Plastic and reconstructive surgery. 1996 Dec ;98(7):1264-71,

discussion 1272-4.

12. Dellon AL, Mackinnon SE. Sciatic nerve regeneration in the rat. Validity of

walking track assessment in the presence of chronic contractures. [Internet].

Microsurgery. 1989 Jan ;10(3):220-5.



Nov ;14(11):1116-22.

14. Mackinnon SE, Hudson AR, Hunter DA. Histologic assessment of nerve

regeneration in the rat. [Internet]. Plastic and reconstructive surgery. 1985 Mar

;75(3):384-8.

15. Almquist E, Eeg-Olofsson O. Sensory-nerve-conduction velocity and two-point

discrimmination in sutured nerves. [Internet]. The Journal of bone and joint

surgery. American volume. 1970 Jun ;52(4):791-6.












nerve. 2002 Nov ;26(5):630-5.




2004 Nov ;21(11):1652-70.

21. Luis A, Amado S, Geuna S, Rodrigues J, Simoes M, Santos JD, et al. Long-term

functional and morphological assessment of a standardized rat sciatic nerve

crush injury with a non-serrated clamp [Internet]. Journal of neuroscience

methods. 2007 ;16392–104.


artificial cells, and artificial organs. 1990 Jan ;18(1):1-24.




[Internet]. Advanced drug delivery reviews. 2004 Jun ;56(10):1467-80.





of America. 1972 Jan ;69(1):258-63.



membranes. [Internet]. Biomaterials. 2005 Feb ;26(5):485-93.




;29(33):4409-19.




137.




Jun ;14(6):979-93.




;18(5):323-8.







Academy of Sciences of the United States of America. 1976 Feb ;73(2):462-6.

32. Laat SW de, Saag PT van der. The plasma membrane as a regulatory site in

growth and differentiation of neuroblastoma cells. [Internet]. International review

of cytology. 1982 Jan ;741-54.



33. Reagan LP, Ye XH, Mir R, DePalo LR, Fluharty SJ. Up-regulation of angiotensin

II receptors by in vitro differentiation of murine N1E-115 neuroblastoma cells.

[Internet]. Molecular pharmacology. 1990 Dec ;38(6):878-86.

34. Larcher JC, Basseville M, Vayssiere JL, Cordeau-Lossouarn L, Croizat B, Gros

F. Growth inhibition of N1E-115 mouse neuroblastoma cells by c-myc or N-myc

antisense oligodeoxynucleotides causes limited differentiation but is not coupled

to neurite formation. [Internet]. Biochemical and biophysical research

communications. 1992 Jun ;185(3):915-24.

35. Prasad KN. Differentiation of neuroblastoma cells: a useful model for

neurobiology and cancer. [Internet]. Biological reviews of the Cambridge

Philosophical Society. 1991 Nov ;66(4):431-51.

36. Prasad KN, Kentroti S, Edwards-Prasad J, Vernadakis A, Imam M, Carvalho E,

et al. Modification of the expression of adenosine 3ʼ,5'-cyclic monophosphate-

induced differentiated functions in neuroblastoma cells by beta-carotene and D-

alpha-tocopheryl succinate. [Internet]. Journal of the American College of

Nutrition. 1994 Jun ;13(3):298-303

37. Yu P, Matloub HS, Sanger JR, Narini P. Gait analysis in rats with peripheral

nerve injury. [Internet]. Muscle & nerve. 2001 Feb ;24(2):231-9.

38. Meek MF, Den Dunnen WF, Schakenraad JM, Robinson PH. Long-term

evaluation of functional nerve recovery after reconstruction with a thin-walled

biodegradable poly (DL-lactide-epsilon-caprolactone) nerve guide, using walking

track analysis and electrostimulation tests. [Internet]. Microsurgery. 1999 Jan

;19(5):247-53.

39. Monte-Raso VV, Barbieri CH, Mazzer N. Índice funcional do ciático nas lesões

por esmagamento do nervo ciático de ratos. Avaliação da reprodutibilidade do

método entre examinadores [Internet]. Acta Ortopédica Brasileira. 2006

;14(3):133-136.

40. Canu M-H, Garnier C, Lepoutre F-X, Falempin M. A 3D analysis of hindlimb

motion during treadmill locomotion in rats after a 14-day episode of simulated

microgravity. [Internet]. Behavioural brain research. 2005 Feb ;157(2):309-21.

41. Canu M-H, Garnier C. A 3D analysis of fore- and hindlimb motion during

overground and ladder walking: comparison of control and unloaded rats.

[Internet]. Experimental neurology. 2009 Jul ;218(1):98-108.

42. Chang Y-H, Auyang AG, Scholz JP, Nichols TR. Whole limb kinematics are

preferentially conserved over individual joint kinematics after peripheral nerve



injury. [Internet]. The Journal of experimental biology. 2009 Nov ;212(Pt

21):3511-21.

43. Lundborg G. Alternatives to autologous nerve grafts. [Internet]. Handchirurgie,

Mikrochirurgie, plastische Chirurgie : Organ der Deutschsprachigen

Arbeitsgemeinschaft für Handchirurgie : Organ der Deutschsprachigen

Arbeitsgemeinschaft für Mikrochirurgie der Peripheren Nerven und Gefässe :

Organ der Vereinigung der Deut. 2004 Feb ;36(1):1-7.



;5293-347.


Chapter 11: Tissue engineering of peripheral nerves [Internet]. Elsevier; 2009.

46. Zacchigna S, Giacca M. Chapter 20: Gene therapy perspectives for nerve repair.

[Internet]. International review of neurobiology. 2009 Jan ;87381-92.

47. Terenghi G, Wiberg M, Kingham PJ. Chapter 21: Use of stem cells for improving

nerve regeneration. [Internet]. International review of neurobiology. 2009 Jan

;87393-403.

48. Dahlin L, Johansson F, Lindwall C, Kanje M. Chapter 28: Future perspective in

peripheral nerve reconstruction. [Internet]. Elsevier; 2009.



50. Perry J. Gait analysis: Normal and Pathological Function. SLACK Incorporated;

1992.

51. Cavanagh P. Biomechanics of distance running [Internet]. Champaign, Illinois:

Human Kinetics Books; 1990.

52. Meek MF, Werff JFA van der, Klok F, Robinson PH, Nicolai JPA, Gramsbergen

A. Functional nerve recovery after bridging a 15 mm gap in rat sciatic nerve with

a biodegradable nerve guide. [Internet]. Scandinavian journal of plastic and

reconstructive surgery and hand surgery / Nordisk plastikkirurgisk forening [and]

Nordisk klubb for handkirurgi. 2003 Jan ;37(5):258-65.

53. Hyatt J-PK, Roy RR, Baldwin KM, Edgerton VR. Nerve activity-independent

regulation of skeletal muscle atrophy: role of MyoD and myogenin in satellite

cells and myonuclei. [Internet]. American journal of physiology. Cell physiology.

2003 Nov ;285(5):C1161-73.



54. Ijkema-Paassen J, Meek MF, Gramsbergen A. Reinnervation of muscles after

transection of the sciatic nerve in adult rats. [Internet]. Muscle & nerve. 2002 Jun

;25(6):891-7.

55. Edgerton VR, Roy RR. How selective is the reinnervation of skeletal muscle

fibers? [Internet]. Muscle & nerve. 2002 Jun ;25(6):765-7.

56. Borisov AB, Dedkov EI, Carlson BM. Abortive myogenesis in denervated skeletal

muscle: differentiative properties of satellite cells, their migration, and block of

terminal differentiation [Internet]. Anatomy and Embryology. 2005 Mar

11;209(4):269-279.

57. Dedkov EI, Kostrominova TY, Borisov AB, Carlson BM. Reparative myogenesis

in long-term denervated skeletal muscles of adult rats results in a reduction of

the satellite cell population. [Internet]. The Anatomical record. 2001 Jun

1;263(2):139-54.

58. Bruusgaard JC, Gundersen K. In vivo time-lapse microscopy reveals no loss of

murine myonuclei during weeks of muscle atrophy. [Internet]. The Journal of

clinical investigation. 2008 Apr ;118(4):1450-7.

59. Cebasek V, Radochová B, Ribaric S, Kubínová L, Erzen I. Nerve injury affects

the capillary supply in rat slow and fast muscles differently. [Internet]. Cell and

tissue research. 2006 Mar ;323(2):305-12.[

60. Huijing P a, Jaspers RT. Adaptation of muscle size and myofascial force

transmission: a review and some new experimental results. [Internet].

Scandinavian journal of medicine & science in sports. 2005 Dec ;15(6):349-80.

61. Meijer HJM, Rijkelijkhuizen JM, Huijing PA. Effects of firing frequency on length-

dependent myofascial force transmission between antagonistic and synergistic

muscle groups. [Internet]. European journal of applied physiology. 2008 Oct

;104(3):501-13.

62. Rijkelijkhuizen JM, Baan GC, Haan a de, Ruiter CJ de, Huijing P a.

Extramuscular myofascial force transmission for in situ rat medial gastrocnemius

and plantaris muscles in progressive stages of dissection. [Internet]. The Journal

of experimental biology. 2005 Jan ;208(Pt 1):129-40.

63. Huijing P a. Muscle as a collagen fiber reinforced composite: a review of force

transmission in muscle and whole limb. [Internet]. Journal of biomechanics. 1999

Apr ;32(4):329-45.

64. Maas H, Meijer HJM, Huijing PA. Intermuscular interaction between synergists in

rat originates from both intermuscular and extramuscular myofascial force

transmission. [Internet]. Cells, tissues, organs. 2005 Jan ;181(1):38-50.



65. Howard CS, Blakeney DC, Medige J, Moy OJ, Peimer C a. Functional

assessment in the rat by ground reaction forces. [Internet]. Journal of

biomechanics. 2000 Jun ;33(6):751-7.

66. Johnson WL, Jindrich DL, Roy RR, Reggie Edgerton V. A three-dimensional

model of the rat hindlimb: musculoskeletal geometry and muscle moment arms.

[Internet]. Journal of biomechanics. 2008 Jan ;41(3):610-9.

functional assessment after peripheral nerve injury · extensor postural, latência do reflexo...

Documents